Home

Master`s Thesis

1. To 2013 11 15 23 59 Y Jump period lt gt gt b Each line presents data for all measuring locations having the same label Figure 5 27 Traffic intensity at A all locations with label EXIT B labels ENTRANCE EXIT Ring_Clk andRing_CoClk for 15 11 2013 Split is made on location are easily distinguishable The last case is when the split is made on combination of vehicle class and location In this case we draw one line for each combination of the selected classes and locations Here each line is painted in dashed style with two colors Figure 5 28 one color comes from the vehicle class and one color from the selected location Here it is also important to select colors that make the chart easy to understand On the figure can be seen that the combination from red EXIT and dark red gt 5 6m lt 12 2m is not a good one 50 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Number of vehicles Chart type Lines Split All From 2013 11 1500 00 To 2013 11 15 23 59 Agg Step l hour Vehicles gt 5 6m lt 12 2m gt 12 2m Locations ENTRANCE EXIT Chart zm Grouping 750 Lines RZ Labels m All split Vehicle Type Locations Split Split M Al Vehicles r m ENTRANCE za v M EXIT
2. 9 00 10 30 12 00 13 30 15 00 5 11 15 11 15 15 11 15 11 Time Chart type Grouping Lines Points O All split Vehicle Type Locations Split Split M Al Vehicles EO01_SRETIO42 E7 lt 5 6 E001_SRETI40 zabawy E001_SRETI67 M gt 5 6m lt 12 2m EO01_SRETIO42r MO gt 12 2m v GEO01_SRETIO43 E001_SRETI072 M Uncategorized Zizcac Al R A OG6O mer Week Days Split Momo Mh MO Average M Tu Miller i u 1 vg mw DOBE Aggregation step 30 min w Unit y Scale a Total Vehicles Automatic Veh Hour Fixed 1639 0 Time Calculation period From 2013 11 15 00 00 To 2013 11 15 raj 23 59 H Jump period lt lt E Chart type Grouping Lines RZ Labels v All split Vehicle Type Locations Split Split WO An Vehicies ENTRANCE v IM EXIT E a lt 5 6m v Ring_Cik W gt 5 6m lt 12 2m W Ring_CoCIk E gt 12 2m M _ Uncategorized 0QG6O Week Days Split M_ v W m M sa El rage ET WF il y Av Ewe Ol Aggregation step 30 min x Unit y Scale Total Vehicles a Automatic Veh Hour O Fixed 8197 0 Time Calculation period From 2013 11 15 00 00 H
3. Bik Ger Aaa F BAI dE e EEA m T gy iun Dhoma ha Awtuuuu nA ho IN pzy Bevel PASA IM caro Hpi aa a ric E 4 ag wt fi 1 mJ En A A A E a m se e li o nn F Te apaiia Le i i Prato a namd acj Laca hepa h Aalto E Bin Z Dr ATU vie w mi r E L R 7 rm erie ok Figure 3 1 Examples of using arrows and bars to present information on a map On image a arrows are used to display the direction of the movement of a single vehicle The thickness of the arrows is proportional to the number of times the vehicle passed through that stretch On image b bars are used to indicate the minimum and median times spent in different places during the trips of the trucks of a delivery company Images are taken from 8 and 5 City Traffic Visualization 7 CHAPTER 3 RELATED WORK Irrespective of the spatial component data that changes over time can present significant challenges to visualization and analysis Currently map animation has become the standard approach to portraying time dependent data and dynamic phenomena 5 Another very popular technique is presenting multiple views of the same map region over various time moments Figure 3 2 Andrienko 4 argues that although these techniques can prove to be very useful in showing the time variances in the data it is still difficult to consider all aspects of the dimension of time in a single visualization
4. 4 A 1558 1673 57 Mi ZE i EF R a eft eos f Veldhoven Figure B 15 Traffis data visualization over a road network Figure B 17 shows the user interface with all available controllers that are used to create and tune these images From combo box 3 the type of visualization can be selected visualizing data over the road network figure B 15 and visualizing data over measuring locations figure B 16 Combo box 1 allows the selection of a color scheme bottom right corner of the image User can select between Partial Hot Body PartialHeat Yellow Orange Red YIOrRd Red 76 City Traffic Visualization APPENDIX B USER MANUAL shades Reds Yellow Green Blue Y GnBl Red Purple RdPu Purple Red PuRd Cyan Magenta CyMa and Blue Green Red BIGnRd Selecting check box 2 makes each color scheme discrete a 5 4 E 4 an Oklenbarne Europalaan Orpheusiaan 4 te 4 k 4 4 4 4 4 12153 25506 4 17314744 nehoef i ie Toren Hondsheuvels MNO p a a a 5 atl 4 4 4 Zwaanstraat w j M mi e D z Eindhoven Beuken a k z E Gildebuurt po A s i b A gt Hemelrijken f cri Limbeek EA Aa 3 EM TE Drents Dorp strijp S A A a A k 4 2 i elolaan 8 1 fessor doctor Darg Limbeek Zuid 5 y p Fellenoord Se A Eindhoven Fra ba a x Fudtla80 Gy 24717 Philipsdorp 18 Septemberplein m PZA 0 w ry a
5. oo 00 H To 2013 01 03 ca 23 59 H A Figure 5 32 charts visualization user interface A Charts type B Visualization type panel C Vehicle type panel D Measuring locations panels E Week days panel F Measuring unit panel G Maximum y Scale panel H Time controllers panel 54 City Traffic Visualization Chapter 6 System architecture In this chapter we describe the architecture of our system and how the separate components are connected together One of our goals is to make things as simple as possible and to automate the process as much as possible hence reducing the extra work for the user City Traffic Visualization Pa f s 2 4 Application gt a NDW historische data A Process Raw Process Data and Fa Data Files Generate Images S Im ages e 7 H e 2 la re lala Administrator j Administrator User Use terminal 4 Raw data F Administrator Fill data Her Prowide feedback Make adjustments Data request Send data Figure 6 1 System architecture and data flow diagram Downloading raw data from NDW is the only process that has to be done manually At the time of writing this thesis NDW historical traffic data servers do not provide an API Application Programming Interface for automating the task of obtaining data The system administrator has to use working station with internet access to log in to the NDW system
6. R9 R10 R11 R12 R13 The application should present the traffic for a city level The data should be presented with the respect to vehicle class User should be able to identify different type of traffic e g incoming traffic outcoming traffic city center traffic etc The application should provide a momentary view of the traffic situation User should be able to compare traffic data with the respect to time vehicle class and location The application has to be interactive allowing the user to make the selections mentioned above The application should not be very complicated to use Users should spend no more than 30 minutes learning how to use it The produced visualization images should be estetically pleasing City Traffic Visualization CHAPTER 2 VISUALIZATION REQUIREMENTS 2 3 Users As we mentioned above the current project has as a goal to develop a useful tool to help the researchers at the Smart Logistics lab in their work This includes professors assistant professors PhD and MSc students All of them are working in the academic field and have background knowledge in the field of logistics and transportation and at least basic skills in working with computer Users can be expected to have experience in reading maps and analytical skills in reading and understanding simple charts The group of possible users can be extended to include other researchers from the OPAC group and other users
7. 11833 Total Height Color Height Figure 5 20 Visualization bar The check boxes are active only during road network data visualization The last two presented panels F and H allow the selection of the measuring unit for the displayed values and the maximum for the color the used color scheme If the option for automatic maximum is selected then the text field displays the current highest calculated value If a fixed maximum is selected then the scheme is adjusted according to the number in the field 42 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Calor scheme Visualization _ Discrete color map Vehicle Type Lo C TAI Vehicles i D lt 5 6m C gt 5 6m lt 12 2m L1 gt 12 2m E Uncategorized MM ETE a een E Wap controllers Markers Values Arrows Map transparency fi Veh Hour m Fixed 1484 Calculation period From 2013 06 24 Ca 00 00 H Jump period Smalljump step 1hur 7 Figure 5 21 Map visualization user interface A Color scheme controllers B Visualization type panel C Vehicle type panel D Measuring locations panel E Map controllers F Measuring unit panel G Maximum panel H Time controllers panel 5 4 Charts visualization Visualizing information with the help of various type charts is a standard approach in pre senting data and giving better understanding Although most of the charts cannot present the spatial property of data
8. 30 min b Unit y Scale m Total Vehicles m Automatic 285 Veh Hour Fixed 468 0 Time Calculation period w From 2013 11 15 00 00 H To 2013 11 15 23 59 475 Jump period lt lt lt gt gt gt Figure 5 30 Bar chart comparing traffic intensity for vehicles longer than 5 6 meters between ENTRANCE and EXIT labeled locations on 15 11 2013 The orange polyline shows the difference between the top and bottom bars 52 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION The second type of charts available in our tool is bar chart By design it shows only one or two lists of values Calculating the values for display is done in the same way as calculating the values for line charts The only difference is that user cannot split the data on any criteria Figure 5 29 shows the daily traffic intensity for vehicles longer than 5 6 meter at ENTRANCE locations The color of the bars is determined by the color assigned to the visualized label data If we choose to show traffic from randomly selected measuring locations then bars are painted in one of the color from scheme 5 24c To make neighboring bars easier to distinguish we gave every other bar a darker shade If the user wants to see the exact value represented in a bar then the mouse cursor should be positioned over it If the user wants to keep seeing the value then he can click on the bar Bar charts allow the option to compare traffic
9. GEO01_ SRETIO42 lt 56m GE001_SRETI40_ PORA I H Ee h SZJ i _ 6E001_SRETI67 s j l gt 5 6m lt 12 2m GEO01_SRETI042r 4 4 gt 12 2m GE001_SRETI043 C Uncategorized A E oD amp i 4 ih OG6GO Mrne lv s a F gt E IA 5490 f CHO Map controllers E q J m O Drents Dorp 7 a 4 a x Ma cj p transparency or Dorge olaan Q Professor doc sy usa Low High Eindhoye p W O eos Unit Maximum RO sdom gt 22 w a Total Vehicles a Automatic omakken Wes Veh Hour O Fixed 26368 332 He 16093 A Eindhoven omakker Wes EE Te Stripses Time ant Calculation period Stes E ASLETTeIN Vi mderk warti By r E eg iN gt From T 00 EJ Hee Z s OL ma 2013 05 24 2 00 00 H Haste aak Zud 56354 Ta z 20324 o 2013 05 2 23 59 5 150 4 16980 Jump period lt lt lt gt gt gt a Rochusbuurt Hagen ide Spoorbaan Small jump step 11 hour er Wurgsty aaf ies 7416156 14606 onskwarti nenpiremn y AR Genneperzijde gt t Bennekel Oost 6 u 3 ai 3 NE ont Hanevoet KS Sa s E Ff roza pores me Y 2 62 gt 0 13184 26368 Figure 5 19 All traffic intensity for the RING in Eindhoven for 24 May 2013 5 3 3 Map visualization user interface Both types of visualization described above can be controlled using the same user interface Thus users have to learn to work with only
10. Git Stratumse Heide A ARAS Figure 3 4 Google Live Traffic snapshot of the traffic in the area of Eindhoven The image shows the traffic speed of the vehicles relative to the speed limit of the roads phones and tablets 16 This allows it to use a large amount of speed data Unfortunately in most of the cases it cannot collect information about the type of vehicle the device is in This makes it impossible to find the difference between the average speed of small cars and big lorries The second drawback of using smartphone data is that Google cannot tell how City Traffic Visualization 9 CHAPTER 3 RELATED WORK many vehicles have passed through a given point as it has information only about the number of Android devises that has passed Another limitation back is the color map that is used to draw the traffic lines It consists of only four colors green orange red and dark red Although it is intuitive for most users to understand green means good and hence in our case fast traffic and for red the opposite the green and the red color seem brighter that the orange and the dark red Hence initially the user s attention is attracted by them and this can lead to making wrong decisions by simply ignoring or not seeing the orange and dark red streets who have slower traffic Also the maps provided have large green areas This can affect the readability if the map especially when a fast traffic road is going through such areas
11. lanes The indexes for the special lines like leftHandTurningLane rightHandTurningLane busLane etc are not used and shown in this thesis This decision was a result of consulta tions with the future users from the OPAC group It turned out that the traffic in such lanes was not of interest to them Another reason is that during this project we did not find any entry in the raw data that has an index different from the ones shown in the figure Flow Speed Flow Speed Flow Speed Flow Speed Flow lengtecategorie Lanes c FC 13 19C 25C 31C 37C 43C gt 1 85 lt 2 4 1 49C gt 2 4 lt 5 6 2C BC 14C 20C 26C 32C 38C AAC 50C gt 5 6 lt 11 5 3C 9C 15C 21C 27C 33C 39C 45C S1C gt 11 5 lt 12 2 C 10C 16C 22C 28C 34C 40C 46C 52C gt 12 2 5C 11C 17C 23C 29C 35C A1C AIC 53C Anyvehicle 6C 12C 18C 24C 30C 36C 42C 48C S4C lt 5 6 1B 5B 9B 13B 17B 21B 25B 29B 33B gt 5 6 lt 12 2 2B GB 10B 14B 18B 22B 26B 30B 34B gt 12 2 3B 7B 11B 15B 19B 23B 27B 31B Etc Anyvehicle AB 8B 12B 16B 20B 24B 28B 32B Geen categorie n 1A 2A 3A AA 5A GA TA BA Figure 4 2 Index values at regular lanes Image is taken from Handleiding Historische gegevens Dictionary NL EN lengtecategorie length category Geen categorie n uncate gorized To ease our work organize the data in a better way and to decrease the time required to find an entry in the raw data we decided to do a second preprocessing step create a new table for each separate measuring location and i
12. 00 00 H Time To 2013 05 19 ra 23 59 H 00 00 02 00 04 00 06 00 08 00 10 00 12 00 14 00 16 00 18 00 20 00 22 00 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 Jump period lt lt lt gt gt gt Figure 5 22 Daily variation of traffic intensity of lt 5 6m vehicle class for all measuring locations with label ENTRANCE for the period from 6 May 2013 to 19 May 2013 Each line shows data for a day Lines are colored depending on the day of the week they show Line charts in our tool present the variation of one or multiple data streams in time Each line in the chart represents a list of values Figure 5 22 show daily traffic intensity data for all ENTRANCE locations combined Each line shows data for a different day All map visualizations described in the previous section allow the user to display only one stream of data This does not provide any possibility to compare different streams and types of data e g we cannot compare traffic intensity between small and medium length vehicles Charts does not have this issue To make comparisons it is important to identify on which property the data has to be compared Our data has three main properties vehicle class location and time Hence the user should be enabled to split the data on any of these properties Splitting the data on the first two is pretty straight forward as there are a fixed number of class
13. DATA PREPROCESSING gegep Ajrsuoqul 9781 ATUO surejuoo 3189 9Y Y BDXH SIN UI peuodo pue MAN Wo paure3qo eyep esl Jo sjdurexo Uy T p 910 314 zaue 09 S6 z zy 3unin suey Jdue I3AyJIUOWIEU o o o S S S0 00 ETOZ TO TO 00 00 TOZ TO TO ZT T OtiLIJYS T0039 z3ue 09 S6 z zy 3unin suey Jdue I3AYJIUOWJEY o o o S S S0 00 TOZ TO TO 00 00 TOZ TO TO SITT Ovbll34S T0039 z3ue 09 S6 z zy 3unin suey Jdue I3AyJIUOWIEU o o o S 5 50 00 TOZ TO TO 00 00 TOZ TO TO SOT T Ovll34S T0039 z3ue 09 S6 z zy 3unin suey Jdue I3AyJIUOWJEU o o o S 5 S0 00 T0Z TO TO 00 00 ETOZ TO TO 86 T Ovll34S5 T0039 Taue 09 S6 z zy 3unin suey Jdue I3AYJIUOWIEU o o o o S0 00 TOZ TO TO 00 00 ETOZ TO TO gp T Ovliaus T0039 T3ue 09 S6 z zy 3unin suey Jdue I3AyJIUOWJEU o o o o S0 00 T0Z TO TO 00 00 ETOZ TO TO ge T Ovbll34S T0039 t uej 09 S6 z zy 3unin suey Jdue I3AYJIUOWIEU o o o o S0 00 T0Z TO TO 00 00 ETOZ TO TO 8zT Ovll34S T0039 Taue 09 S6 z zy 3unin suey Jdue I3AyJIUOWJEU o o o o S0 00 TOZ TO TO 00 00 TOZ TO TO att Ovliaus T0039 Taue 09 S6 E 33M125 EY Jdue I3AyJIUOWJEU ra o vz 5 S S0 00 T0Z TO TO 00 00 ETOZ TO TO Sb T ZYOLIJ4S T0039 t uej 09 S6 JE 332M12 5 EY idue I3AyJIUOWJEU o o o S S S0 00 T0Z TO TO 00 00 ETOZ TO TO ge T TtOLLIYS T0039 T3ue 09 S6 T Samuays ey Jdue I3AYJIUOWJEY o o o S S S0 00 T0Z TO TO 00 00 ETOZ TO TO 8zT ZYOLI34S T0039 Taue 09 S6 T 3 mvu 345s ey Jdue I3AyJIUOWJEU ra o bz 5 S 50 00 TOZ TO TO 0
14. Using their provided web interface the administrator has to specify what data he wants to download After the request has been processed the administrator can download the data http ndw dysi nl City Traffic Visualization 55 CHAPTER 6 SYSTEM ARCHITECTURE back to his working station Here we have to note that with the use of the term system administrator we mean any user who has the privileges and credentials to access and work with NIDW servers The city traffic visualization system has two main components a data storage module database and data processing module software application The data storage component is used to store the processed traffic data described in Chapter 4 In our final implementation we set up a PostgreSQL database for our data storage module The original implementation was done using csv files to store the data Each table described in Chapter 4 is saved in a separate file The main advantage is that it is simpler to set up as the user does not have to install and tune a database and also that the files are stored directly on the users machine hence always available Unfortunately the increasing size of the files leads to a decrease in performance of our application Because of that we considered a database approach A few speed tests using both a file and a database storage system were done We processed two different size chunks of raw data the first one containing traffic data from all measuring locatio
15. d postgres H Catalogs 2 M Event Triggers 0 ci Extensions 2 H Schemas Refresh 2 Tablespaces 2 New Extension 42 Group Roles 0 4_ Login Roles 2 Object List Report Add all postgis extensions For more information visit also http www bostongis com printerfriendly aspx content_name postgis_tut01 New Extension Properties Definition SQL Name pag_freespacemap OID pag_stat_statements _trgm pgcrypto pgrouting pgrowlocks pgstattuple pidbgapi 9 Set up Login Roles on the server A login role is used by the application to connect and work with the database 10 Create a connection string text file The file should have three lines First jdbc postgresql Host port DatabaseName Second and third line contain Username and Password 62 City Traffic Visualization APPENDIX A ADMINISTRATOR MANUAL 3 ConnectionString txt Notepad File Edit Format View Help jdbc postgresql localhost 5432 TrafficData user 123456 A 2 Download and import data from NDW to the database There are several types of files provided by NDW In our application we used only RVM network data file and traffic intensity data files The first one has to be imported manually into the database the second type of files are imported with the help of our tool 1 Go to NDW website http www ndw nu gt Databank Actuele verkeersgegevens or click http www ndw nu pagina n1 4 d
16. lt 5 6 600 E i Wi gt 5 6m lt 12 2m Wr gt 12 2m M _ Uncategorized 450 Week Days Split ia m ksi ie i 300 Aggregation step 11 hour v Unit y Scale Total Vehicles Automatic Veh Hour O Fixed 6530 Time 150 Calculation period From 2013 11 15 00 00 H To 2013 11 15 23 59 HH Jump period lt lt lt gt gt gt Time 0 12 00 15 11 Figure 5 28 Traffic intensity for gt 5 6m lt 12 2m and gt 12 2m vehicle classes at all locations with label ENTRANCE and EXIT for 15 11 2013 Split is made on vehicle class type and location Each polyline shows data for one vehicle class and one location The proposed color schemes are used by default but the user can always change the colors assigned to the vehicle classes week days and labels To increase the readability of the chart and to aid data analysis some user interaction is implemented If the user positions the mouse cursor over a line this will result in drawing the line highlighted thicker line with black outlines and also to be positioned on top of the other lines A small box is drawn showing information about the line Depending on the selected split information can vary between the vehicle class that is represented the day of the week or the location name label If the cursor is positioned over a value point then the value is shown above the point Figure 5 26 City Traffic Visualiz
17. provides data only for traffic going in the same direction This is important information to be taken into account Hence upon binding location to a road the user should also provide information about the direction of the traffic that is measured Figure 5 4 We implemented a semi automated function which can save the user some time More details about it can be found in Appendix B The last piece of information that is specified in the model is the information about labels 24 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION A B Figure 5 2 A Shortest distance from point M to line a can be found by dropping a perpen dicular line from M to a and measure its distance d B Actual image taken from our tool presenting the distance between three measuring locations 1 2 3 and two roads A B We know for fact that point 1 should be positioned on road B while points 2 and 3 on road A Green lines show the shortest distance between the points and any road The red line shows the distance between point 2 and road A Positioning locations using shortest distance to a road would result an incorrect positioning of point 2 on road B The green and the red lines are only annotations and are not part of our application of measuring locations Requirement R8 states that user should be able to identify different type of traffic i e there must be some method for grouping measuring locations with similar properties This is d
18. search and fetch the needed information Visualizing all available data in one image is practically impossible Hence we designed a number of views that address various properties of the data in more details Visualizing traffic data over a road map allows the user to get a spatial view of the traffic status Analyzing this type of images make it fast and easy to identify the main traffic arteries and their state during the different periods of the day or year A number of charts has been developed for visualizing the traffic They can present the time variance of the data in detail Researchers can use them to compare the traffic in different locations or compare the traffic for different time periods The user is provided with the ability to select and filter the data which he wants to see We followed some of the advices from related work to create clear and meaningful visualizations As our data can be explored in various ways and the usefulness of the visualization depends on the research questions we leave it to the user to make adjustments and we provided several options to control time periods chart types colors and other map properties These encourage users to explore the data with our tool Our work has been focused in visualizing NDW data and we optimized our techniques to show its properties Therefore our approach might not be applicable to other types of data The developed system is divided into a number of modules Using a data
19. they can show temporal aspects in small detail As most of the researchers in our target group need to answer questions concerning time we included this type of visualization in our tool Being one of the most common type of charts line and bar charts were adapted to suit our needs Most of our users are experienced in working and reading these type of charts and we believe that this will help our tool being easier to learn City Traffic Visualization 43 CHAPTER 5 DATA VISUALIZATION and use Both types of charts require a built city model to use their full functionality 5 4 1 Line charts data visualization Number of vehicles Chart type Lines Split Time From 2013 05 06 00 00 To 2013 05 19 23 59 Agg Step l how Vehicles lt 5 6m z zr Week Days Mo Tu We Th Fr Sa Su Locations ENTRANCE Chart type Grouping 17500 Lines w Labels v All split Vehicle Type Locations Split Split WC All Vehicles M ENTRANCE EXIT 14000 M lt 5 6m m M gt 5 6m lt 12 2m M _ gt 12 2m M _ Uncategorized MM EA A 10500 Week Days Split Hmo M y Tn Mily Sa Average MelTu M vjrr Mv Su _ Only Avg 7000 Enw OORO Aggregation step 1 hour v Unit y Scale a Total Vehicles Automatic 3500 Veh Hour Fixed 15701 0 Time Calculation period From 2013 05 06
20. vehicle class on a section of Boschdijk street The red double arrows show the possible entrance exit locations for the traffic on the top lane going left between measuring locations A and B 5 3 2 Visualization of traffic at the measuring locations The next view which was created for presenting data on a map allows visualization of the collected data on the locations where it was collected Figure 5 19 Instead of using the calculated values for coloring the road network here it is used to draw a small bar which parameters depend on the visualized value The bars are used in a separate view because we want to use colors to present the values at the exact location where the measurement took place Using colored bars over colored roads would result in a messy image Calculating the values and mapping them to colors is done the same way as described in the previous section All visualization bars have the same size defined by black outlines Figure 5 20 This makes adjacent bars comparable to each other The inner solid rectangle colored orange is used to visualize a value Its height and color are variable and depend on the value that is visualized The color depends on the selected color scheme and if a maximum displaying value is set The height is calculated via l cH eight va x tHeight maa if cHeight gt tHeight 5 4 then cHeight tHeight where cHeight is the height of the inner rectangle val is the visualized
21. B r Fl e LARE ad eae pperiodEnd computatiomMethod AJ measurementEquipmentTypeUsed AK Ss V A AB period SSS AC specifilane oo Table 4 2 List of the column names that can be found in PointsMetaData table AB Apart from the meta data column we also kept the columns periodStart periodEnd and added indexes The first two keep the timestamps of the first and the last entry for the corresponding measuring location i e we know the period for which we have measurements The column indexes keeps all the indexes that are used for that location e g if we consult with Figure 4 1 we can conclude that the indexes field column C for location GEO01_SRETIO43r will have value 1B 2B 3B 4B The indexes for each location depend on number of lanes of the road there and on the number and on the number of vehicle classes that are recognized by the measuring hardware Similar to the meta data they are constant for every measuring point Index values encode more information about the measurement itself They can give infor mation about the location of the measurement which lane of the road the type of measure ment flow of vehicles or speed and the type of vehicles for which the measurement is made vehicles are divided into groups by their length Figure 4 2 shows the indexes for regular City Traffic Visualization 17 CHAPTER 4 DATA PREPROCESSING
22. One big advantage of Google Live Traffic is that it keeps history of the traffic data and can recreate the traffic speed at any past time upon user request Another application that shows traffic congestions is Onderweg It is developed by one of the NDW partners and uses NDW traffic data It uses all the available traffic speed data and uses it to create image of the current traffic situation Figure 3 5 sa s a geen lch E regulier M zwaar Mi extreem a M ja u T Figure 3 5 Snapshot of the congestions in low mid part of Netherlands made using Onderweg The image shows the relative size of the traffic congestions Unfortunately NDW provides very little traffic speed data which limits the usability of the application At present such data is available only for some of the major highways An advantage of this tool is that it can provide information about current roadworks and hence this information can be used by the user to understand why some roads have congestions Onderweg suffers from the same shortcomings as Google Live Traffic Firstly the colormap is badly designed as the yellow color is very bright and at some places it merges with the gray background used Secondly even though traffic speed data is available per vehicle class short vehicles medium length vehicles long vehicles etc it seems that the application does not http www onderweg nl 10 City Traffic Visualization CHAPTER 3 RELATED WOR
23. S6 T ZV 9W Jdue I3AyJIUOWJEY o o o t v 50 00 TOZ TO TO 00 00 ETOZ TO TO az T 1 vOLI34S T0039 Taue 09 S6 T ZV 9W idue I3AyJIUOWJEY ST o 186 SZ t v 50 00 TOZ TO TO 00 00 ETOZ TO TO att 1 vOLI34S T0039 ngideds poled 021n33 Nseaw JNSEJ3W JNSEJU 31nse3uu 15jUu3W3JNSE3W 13JNSE3LU juonqejndwoj 3 3ne aAdne aAdne 211 1ep nu js DS ns JJepue s 03 Jnulw isqtunu sqiunu pugpon d pezsponad x3pU Se34u 43 15 USLU3JNSE3LU JW av A n T D d O NW1IMA f 3 a 7 Vv 6z 8z Lt 9z s v EZ ce TE oz 6T 8T T 9T TT 15 City Traffic Visualization CHAPTER 4 DATA PREPROCESSING excel column column tle odl column column KONW KONNA KONA WE ER S E B measurementSiteVersion Z measurementSide e AM acemaey O yO DO porod aB eo O O perioden JAG specifilane OOO mumberOffncompletelnputs AD specificVehieleCharacteristics mumberOfluputValucsused ABT statLocatieFoeDiEplaylat 7 minutesUsed ARO start LocatieMonDisplayLong I computationalMethod AG LocationCountryCode standaraDeviation an LocatiowTableNumber supplierCaleulatcDataQuality A1 LocationTableVersion ODO Low ADO alertODirectionCoded oom A pao number ofsGDQ a ofetDistanee O wn JAM OCA travelTimeType an OO DES O O angelica fao ROADNUMBER R aveVehiclespea APO ROADNAMB avg TravelTime AQ FIRSTNAME TT computationMethod ARO SECOND NAME U meaeuromentEquipmentTypelsed AS messageType W mea
24. U Testing the Usability of Interactive Maps in CommonGIS Cartography and Geographic Infor mation Science pp 325 342 2002 8 Andrienko G Andrienko N Bak P Keim D Wrobel S Visual Analytics of Move ment Springer Heidelberg New York Dordrecht London 2013 7 9 Andrienko N Andrienko G Exploratory Analysis of Spatial and Temporal Data Springer Germany 2006 10 Colorbrewer 2 0 Color advice for cartography http colorbrewer2 org 11 Few S Table and Graph Design for Enlightening Communication perceptual Edge 8 12 Few S Effectively Communicating Numbers Selecting the Best Means and Manner of Display November 2005 perceptual Edge 8 13 Few S Show Me the Numbers Designing Tables amp Graphs to Enlighten Analytics Press Oakland 2012 8 14 Few S Edge P Line Graphs and Irregular Intervals An Incompatable Partnership Visual Business Inteligence Newsletter November December 2008 8 City Traffic Visualization 99 BIBLIOGRAPHY 15 Few S Edge P Practical Rules for Using Color in Charts Visual Business Inteligence Newsletter February 2008 116 Google Live Traffic http en wikipedia org wiki Google_Traffic 9 17 INRIX Inrix Traffic http www inrixtraffic com 18 Jankowski P Andrenko N Andrenko G Map centered Exploratory Approach to Multiple Criteria Spatial Decision aking International Journal of Geographical Informa tion Science pp 100 127
25. can easily distinguish between shades of the same color than between shades of different colors Most of the schemes which we designed and used are inspired by ColorBrewer This website gives many discrete color schemes specifically designed to be used on geographical maps The schemes suggested there depend on the type of data visualized sequential diverging and qualitative We need to visualize sequential data here Schemes for sequential data are given in discrete form with up to 9 colors To make a continuous scheme out of them we use a 9 color scheme as a guideline and do a smooth transition between them The result can be seen on Figures 5 12b 5 12c 5 12d 5 12e and 5 12f On Figure 5 12a the partial hot body color scheme is shown The hot body scheme is relatively well accepted within the visualization community It provides smooth transitions between the neighboring shades without many sudden jumps in the colors The scheme goes from white to yellow orange red dark red and black For our needs we dropped the white and black color white because it is widely used in the background map and black because it is too dark a color and attracts user attention less than the red color shades Thus we ended up with a color scheme very similar to Yellow Orange Red scheme by ColorBrewer The main difference is that our colors look brighter The next scheme designed by us is the Cyan Magenta scheme Figure 5 12g It progres sively changes color fr
26. coef ficients adjusted accordingly to RGB components brightness were used Drawing measuring locations markers values and arrows on the image can make it clut tered and difficult to read Because of that we presented the user with an option to select if he wants them or not Figure 5 17 There is one major issue with our visualization approach illustrated on figure 5 18 If we focus on the top lane left going traffic on the section between measuring locations A and B we can see that this street is uniformly colored The next color change happens after location B darker orange indicating that traffic intensity is larger One can make the wrong conclusion that the traffic between locations A and B has the same intensity for the whole stretch It is wrong because we color the street according to the traffic measured in location A which is at the beginning of the section The red arrows indicate possible entrance and exit locations for the traffic moving on the top lane As can be seen on the figure all these entry exit locations result in increased traffic intensity on Boschdijk in location B Hence users should keep in mind that a stretch from a road is colored according to the measured traffic at its beginning and the result of all vehicle movement on that stretch can be seen on the next measuring location City Traffic Visualization 39 CHAPTER 5 DATA VISUALIZATION Figure 5 18 Traffic intensity of gt 5 6m lt 12 2m
27. e Sub et Bod PRUE GE o aa I Bai Ade ees ea ecko 3 22 OJUMOMIENE e Ed ie a a WAWA a d M Ko Gd a a i 4 Dies A O Ge e Coe Sw A Bode e e as ft ny ee Nar Ge dei a e Te es 5 3 Related Work T 4 Data Preprocessing 13 4 1 Nationale Databank Wegverkeersgegevens NDW 13 A Trane Data 2 a gaa ae ee he ed a es Ak hes ee Wam GE M a ae 14 A Roads Network Data s e gae e ie a a a za es ee ZZ 20 BA Omer Daa O oe Bee Z 22 5 Data Visualization 23 xt Ci Modeb a 2250 a AA AA BOS So 23 SQ Basic contr ll rs s s e ea a a tia e a de Ge A 28 93 Visualization ol traici n a Map 223 24 6 dns e dd a 29 5 3 1 Visualization of traffic over the road network 30 5 3 2 Visualization of traffic at the measuring locations 40 5 3 3 Map visualization user interface 0 002 002 2 1 Al Dz Charrewi nalzali m ovaw e wa ceo ae de w se we See a ee A RZ ee de ts ta 43 5 4 1 Line charts data visualization 0 0800 bee eee 44 54 2 Barchartsdata visualization s lt a sewa e686 bse kawki ba ews 52 5 4 3 Charts visualization user interface 2 08084 53 6 System architecture 55 7 Conclusions 57 oll Fute Woke NE 58 City Traffic Visualization vii CONTENTS Bibliography Appendix A Administrator Manual B User Manual vili 59 61 61 66 City Traffic Visualization Chapter 1 Introduction The rise of e commerce means an increase of delivery trucks driving arou
28. encoded In a more thorough analysis it is important to know the numbers that are visualized This is done by displaying the value which is used for coloring the road just above the measuring locations marker One small feature which was added upon user request is to draw small arrows on the roads pointing the direction of traffic The arrows are drawn on top of the parallel polylines which we color according to traffic intensity For every section AB of 38 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION 13153 15530 i A Figure 5 17 Showing the visualized values left and drawn arrows on the roads right polyline P an arrow is added on the location of point B pointing in direction AB To make arrows more visible their color is determined based on the brightness of the color of the street they are drawn over Ifit is a light color then arrows are drawn in dark brown and if the color is dark then arrows are in light gray Colors are selected to be in good contrast with the colors used for the streets br Back 0 213 x back R 0 715 x backG 0 072 x back B if br Back lt 0 5 then Background is DARK else Background is BRIGHT 5 3 where backR backG and backB are the Red Green and Blue components of the color in the range 0 1 The brightness of the background color is determined by summing its RGB components Because the green component increase color brightness more than red and blue weight
29. extension pack 1 Download the latest version of PotgreSQL from http www postgresql org download Make sure that you got the package suitable for your operating system 2 Run the downloaded file and install PostgreSQL 3 During the installation use StackBuilder to include the PostGIS add on package 4 Optional PostGIS included in StackBuilder might not be the latest version We rec ommend to get and install the latest PostGIS from http postgis net install 5 The installer for PostgreSQL includes also pgAdmin III http www pgadmin org a graphical tool for managing and developing databases Locate pgAdmin III and open it 6 Create a new server or add an existing one by clicking File gt Add Server Fill the required information and press OK For more information and instructions see the pgAdmin III documentation www pgadmin org docs dev index html or Post ereSQL wiki page https wiki postgresql org wiki Main_Page 7 Create a new database in the server by right clicking on Databases in the server menu and select New Database City Traffic Visualization 61 APPENDIX A ADMINISTRATOR MANUAL File Edit Plugins View Tools Help LAAM ki NIE ad Object browser Server Groups E Z Servers 4 J ExampleServer localhost 5432 8 Add PostGIS extensions to the database Right click on Extensions option of the database and select New Extension
30. nine colors More colors makes neighboring shades difficult to distinguish ColorBrewer also propose discrete schemes with at most nine colors User is allowed the option to manually set the maximum value which is associated with the darkest most right color on our color schemes Colors are mapped evenly to values between 0 and set maximum If later we need to obtain a color for value greater than the set maximum then this values is mapped to the last color of the scheme Figure 5 16 presents traffic intensity for Eindhoven for a whole day divided into periods of 4 hours We can see how traffic increases in the morning and decreases in the evening and also see intensity variations for individual streets during the day Fixing the maximum value can be used to compare traffic hourly daily weekly City Traffic Visualization 37 CHAPTER 5 DATA VISUALIZATION aa i STWA 2 a y i i TA DA ADA Woensel N gt pid 3 2 GFT A i Woensel Nh Woensel l i 49 A e A y H MAS 4 P 7 t y A A e amp e i sal I K y w A al 2 5 E JA Ne r AA j A 3 x y 7 RN T a A Y EN X EB sa e RYSA ia Ak b RYJA e Pe ox ge Woensel Zuid A oa e e NO of Woensel Zuig eee Bar 3 58 Ta lt a AB 1 584 mie QUES AS at el m CER EAR ea w gt NA Fa aa RE z W gi AP NB o WE F T se z n a A gt GR 2 p NY DE ANA FL R Nf ew fra We na 4 NS i d A
31. only on the central part of Eindhoven especially the ring and the roads that lead to it from outside We have removed the unnecessary roads and the roads for which we have no measuring locations and the locations which are not in the city Immediately we can notice a more clear picture because of the reduced visual clutter caused by drawing unnecessary road lines Because we want to visualize data for a whole road rather than a single location we need to identify which measuring location relates to which road of the RVM network NDW does not provide us with clear information for this Measuring locations are given with their geographical coordinates and name of the street on which they are located but the roads are given only with their coordinates and not a name We tried to automate the process using the shortest distance from a measuring location to a road but this gave us some issues City Traffic Visualization 23 CHAPTER 5 DATA VISUALIZATION Wiz MW LI ELE BRT a E peer red A 5 ie t iras B Trimmed Data Figure 5 1 Model of the city of Eindhoven Image A displays all the available information which we have about road network and measuring points locations Image B shows only the information which is needed to create a sample model of Eindhoven Black lines show the RVM road network and the yellow dots represent the location of each measuring station Measuring locations are positioned using the coordinates that comes
32. or any combination of these Filter based on the completeness availability of the measurements For a full description of the user interface consult the Handleiding Historische gegevens available on the NDW website The selected data is available for download in Comma Separated Values csv files This allows us to open them with a text editor or spreadsheet application like Excel or similar and explore the content see Figure 4 1 We can see that the provided data has several attributes Table 4 1 The fields that are highlighted in blue represent the meta data for each measuring location showing the physical location of each measuring point and also the measuring equipment that is used This meta data stays constant it does not change with time and does not depend on the result of each measurement Each location has a unique name measurementSiteReference If some measuring location has to be moved NDW does not correct the location fields from V to AV but instead creates a new entry in their database with a new unique name and new updated meta data Hence we assume that once one measuring location is put into the database its meta data does not change In this assumption we exclude the case where the entry is edited because of some technical mistakes like a spelling error or wrong data in the field http www ndw nu pagina nl 4 databank 65 historische_gegevens 14 City Traffic Visualization CHAPTER 4
33. or double click with the left mouse button to zoom in on that location Positioning the mouse cursor over any measuring location marker pops up an information box presenting most of the meta data for this location Removing the cursor from the marker causes the box to disappear 82 City Traffic Visualization
34. respectively the end date and hour The small calendar button next to the date boxes gives the option to make a 18 City Traffic Visualization APPENDIX B USER MANUAL fast date selection using a calendar menu Button group 19 make quick jumps forward and backward in time 19a gives a jump back in time with a step equal to the selected time interval length From Time ToTime 19d makes the same step but forward 19b and 19c make steps backward and forward in time but with the step defined in combo box 20 Number of vehicles Chart type Lines Split Time From 2013 05 06 00 00 To 2013 05 19 23 59 Agg Step l hour Vehicles lt 5 6m Week Days Mo Tu We Th Fr Sa Su Locations ENTRANCE 17500 14000 10500 7000 3500 ma Time 02 04 00 06 00 08 00 10 00 12 00 14 00 16 00 18 00 20 00 22 00 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 09 05 Figure B 18 Traffic data visualization with the use of line chart Charts visualization allows visualizing data with line or bar charts The user interface figure B 20 is similar to the one used for map visualizations Most of the controllers are the same and have the same function Here only the new controllers are explained Check box 1 allows the selection of the type of chart lines figure B 18 or bars fig ure B 19 With check box 3 user can select if he wants to vis
35. selected time interval Width 1 if space gt label Width then Put label below point pj space 7 1 7 1 n 7 2 n 5 5 Where n is the number of values in a list 2 is the index of the last point p which has a label 46 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Number of vehicles Chart type Lines Split Vehicle From 2013 05 24 00 00 To 2013 05 24 23 59 Agg Step l hour Vehicles All Vehicles lt 5 6m gt 5 6m lt 12 2m gt 12 2m Uncategorized Locations ENTRANCE 17500 c l i l l l 14000 10500 yHeight 7000 3500 a RR Bae M Time Se ee ooo TET 01 00 Eri 00 03 00 04 00 05 00 06 00 07 00 08 00 09 00 10 00 11 00 12 00 13 00 14 00 Sher FET 17 00 sa 19 00 20 00 21 00 22 00 23 00 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 24 05 Figure 5 23 Traffic intensity at all locations with label ENTRANCE for 24 05 2014 Split is made on vehicle class type Green line represents All vehicles class blue lt 5 6m dark red gt 5 6m lt 12 2m purple gt 12 2m and cyan gt 5 6m lt 12 2m 1Width and yHeight show the size of the working plot used for drawing the visualization polylines and Width is the width of the drawing plot Putting label
36. selection of a time interval that is used to aggregate the data The aggregation step also determines the jump period of the lt and gt button of time panel H Every selection made by the user is shown in a small legend on the top of each chart It can be used firstly to understand what is shown on the image but also to recreate it if needed in the future Chart type Lines 5pht Tone From 2013 01 03 00 00 To 2013 01 05 33 59 AgeStep l hour Velucles gt 5 6m lt 12m 12 2m Week Days Mo Tu We Th Fr Locations ENTRANCE EXIT a Example of legend in lines type chart Chart type Bars pht None From 2013 01 03 00 00 To 2013 01 03 23 59 Agg Step 1 how Velucles gt 5 6m lt 12 2m 12 2m Locations ENTRANCE EXIT Compared Compare Line b Example of legend in bar type chart Figure 5 31 Examples of line and bar chart legends City Traffic Visualization 53 CHAPTER 5 DATA VISUALIZATION Chart type Grouping p L 0 All split Vehicle Type Locations Locations a Split 3 Split Split MI An Vehicles HIM ENTRANCE gt E ENTRANCE M_ lt 5 6m m lx Compare Mv gt 5 6m lt 12 2m M 12 2m M Uncategorized Compare Line M mo Wm M Sa Average M r Mi Ir W su _ Only Avg av 9085 F Aggregation step ihour sd G Unit y Scale m Total Vehicles m Automatic H O Veh Hour Fixed 4280 Ime Calculation period From 2013 01 03
37. then each segment AB of the polyline P that describes the main road is taken Figure 5 9 30 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION 4 Color scheme Visualization a eree PartialHeat w Roads RZ j Discrete color map ELU f GMA 7 Vehicle Type z j Y Sx Y All Vehicles GE001_SRETI042 a j A i of gt H 2 A pa lt 5 6m GEO01_SRETI40 Woensel No F Woensel Noorc JGE001_SRETI67 gt 5 6m lt 12 2m GEO01_SRETI042r gt 12 2m GEO01_SRETI043 Uncategorized Q a E cues Map controllers oe Vv Markers v Values C Arrows Map transparency Q 1 uy Low High 1 4 0 am Unit Maximum Total Vehicles a Automatic Veh Hour Fixed 39361 668 Time a Calculation period From 2013 05 24 00 00 H To 2013 05 24 23 59 H v gt gt Jump period lt lt lt Small jump step 1 hour 2279 portret 4 Veldhoven f o 17164 r Co Figure 5 8 All traffic intensity for the city of Eindhoven for 24 May 2013 The image is created with our tool Then we find points A A B B such that AA AA BB and BB AB and AA AA BB and BB d This ensures that A B and A B are parallel to AB have the same
38. value maz is the maximum value and tHeight is the total height of the Visualization bar Drawing the colored bar inside a white rectangle with black borders has several advan tages Firstly the white rectangle provides the same background for all visualized values This eliminates some optic illusions that occur with identifying a color when it is drawn on different backgrounds Hence colors that are used for visualization of similar values should be recognized as almost the same Secondly as color is used to encode a value the height of the colored rectangle also encodes the value Knowing that the total height of the visualization bar is linked to the maximum visualized value or manually set maximum the user can see clearly how close this value is to the maximal one Because it is difficult to determine the exact value from the shades of the color scheme or the height of the inner bar a number displaying the exact value was added above each visualization bar Each bar is connected to its corresponding measuring location Figure 5 20 40 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Color scheme Visualization Bar PartialHeat D Locations v 1 E 4 4 lt _ Discrete color map f Vehicle Type Locations Europalaan Grpheusiaan Z AIl Vehicles
39. 0 00 ETOZ TO TO BIT THOLLIYS T0039 zaue 09 S6 z TW L9W sn I3AYJIUOWIEY T T o S0 00 TOZ TO TO 00 00 TOZ TO TO IST T 1 TLI34S T0039 zaue 09 S6 z ZW 9W sn I3AyJIUOWJEU T T o G0 00 ETOZ TO TO 00 00 ETOZ TO TO DLT T 1 TLI34S T0039 zaue 09 S6 z TW L9W sn I3AyJIUOWJEU T T o G0 00 T0Z TO TO 00 00 TOZ TO TO J9TT 1 TLI34S T0039 zaue 09 S6 z TW L9W sn I3AyJIUOWIEY T T o G0 00 TOZ TO TO 00 00 TOZ TO TO ST T JETILISS T0039 zaue 09 S6 z TW L9W sn I3AyJIUOWIEU T T o G0 00 T0Z TO TO 00 00 ETOZ TO TO IT T 1 TLI34S T0039 zaue 09 S6 z TW L9W sn IBAYJIUOWIEU T T o G0 00 TOZ TO TO 00 00 ETOZ TO TO IET T JETILLIYS T0039 Taue 09 S6 z TW L9W sn I3AyJIUOWJEU T T o S0 00 TOZ TO TO 00 00 ETOZ TO TO 29 T JETILLIYS T0039 Taue 09 S6 z TW L9W sn I3AyJIUOWIEU T T o S0 00 TOZ TO TO 00 00 ETOZ TO TO 95 T JETILLIYS T0039 T3ue 09 S6 z TW L9W sn I3AyJIUOWIEY T T o S0 00 T0Z TO TO 00 00 ETOZ TO TO Jv T 1 TII34S T0039 Taue 09 S6 z TW L9W sn I3AyJIUOWIEU T T o G0 00 T0Z TO TO 00 00 TOZ TO TO DE T JETILLIYS T0039 Taue 09 S6 z TW L9W SN IBAyJIUOWIEU T T o S0 00 T0Z TO TO 00 00 ETOZ TO TO 97 T JETLLIYS T0039 Taue 09 S6 z TW L9W sn IBAYJIUOWIEU T T o S0 00 T0Z TO TO 00 00 ETOZ TO TO TT JETILIWS T0039 T3ue 09 S6 E TW L9W Jdue I3AyJIUOWJEY 09 o 9zv Ch t S S0 00 ETOZ TO TO 00 00 ETOZ TO TO S T 1 vOL134S5 T0039 Taue 09 S6 T Zv 9W Jdue I3AYJIUOWIEY o o o t v 50 00 TOZ TO TO 00 00 ETOZ TO TO SE T 1 vOL134S T0039 Taue 09
40. 10 3070 Itype MultiLineStrinq coordinates 4 51583459437823 3 100 3990 Itype MultiLineString coordinates 4 793769419557 5 Figure 4 6 tmc_line table The final result after performing Step 3 4 4 Other Data NDW provides two more pieces of information again stored in the form of Shape files The first one contains the outlines of the municipalities in Netherlands gemeente in Dutch the second piece the coordinates of all measuring locations Both Shapefiles have the same form as the one presented in the above section The municipality outlines are given as a sequence of coordinates while the locations are given with only one set of coordinates Although both files provide us with usable information we decided not to use it for the following reasons Municipality outlines can be useful but we do not need them for something more than just showing them on the map In our application we use as background a map on which they are already drawn hence we can avoid using this file The locations file gives us only the coordinates of the locations and no other useful information Hence we prefer to construct the PointsMetaData table described in Section 4 2 Table 4 2 22 City Traffic Visualization Chapter 5 Data Visualization In Chapter 2 was stated why it is important to have a good visualization of city traffic data Most of the existing tools and applications do not provide the full functionality which is required for th
41. 2001 19 Karnick P Cline D Jeschke S Razdan A Wonka P Route Visualization using Detail Lenses In IEEE Transactions on Visualization and Computer Graphics pp 235 247 March April 2010 9 20 Salvatore R Pedreschi D Nanni M Giannotti F Angrenko N Andrenko G Visually Driven Analysis of Movement Data by Progressive Clustering In Information Visualization pp 225 239 September 2008 21 Smart Logistics Smart Mobility Intelligent and Productive Mobility and Transport https www youtube com watch v rC7JFijh_YI 3 22 Wijk J Selow E Cluster and Calendar based Visualization of Time Series Data In Information Visualization 1999 Info Vis 99 Proceedings 1999 IEEE Symposium on pp 4 9 October 1999 58 23 Willems N Visualization of Vessel Traffic Ph D thesis Eindhoven University of Tech nology 2011 9 60 City Traffic Visualization Appendix A Administrator Manual This manual is intended to help system administrators set up the database fill up the raw data and prepare everything for work Some of the steps here are optional and can be skipped or done using a different approach What we describe here is an exemplary manual for getting started A 1 Installing and setting up a database The first thing to do is to download and install the required software For our implementation we choose to work with the PostgreSQL object relational database management system with PostGIS
42. APPENDIX B USER MANUAL points on the image figure B 15 box 12 enables the small rectangles with values above the markers Box 13 shows arrows over each road pointing at the direction of the traffic Slider 14 adjust the transparency of the background map With radio buttons 15a and 15b the measuring unit of the visualized data can be selected Total number of vehicles or Vehicles per hour Radio buttons 16 allow you to select the maximum value used for the color scheme 16a makes automatic selection and 16b allows you to put it manually The value in text box 16b1 is used Ly Color scheme Visualization PartialHeat _ gt Roads 3 3 2 Discrete color map Vehicle Type Locations 4a All Vehicles GE001_SRETI042 gt 4b lt 5 6m _ GE00 Ac gt 5 6m lt 12 2m Ad gt 12 2m de Uncategorized A Ls Se 10 NEW LABEL w L6 7B Map controllers Markers Values _ Arrows 11 12 13 14 Map transparency Low High Unit 163 Maximum 138 Total Vehicles a Automatic 15h Veh Hour Time 17a Calculation period a a e 18a From 2013 01 03 a 00 00 H j Toe 2013 01 03 ra 23 59 H 18b 19a 19b 19c Z Jump period ee lt gt gt 20 Small jump step 1 hour w Figure B 17 Map visualizations user interface Time controllers 17 20 allow the selection of a time period for which data is visualized Boxes 17a and 18a show the start date and hour and 17b and 18b
43. ETIO12r Uncategorized m da Jump period ENTRANCE C A B Figure 5 6 User interface for applying basic filtering A Filter for vehicle type B Filter for measuring locations C Filtering based on time period We tried to design the interface to be simple easy to use and time saving All controllers that are responsible for one filter are put together in a separate panel Figure 5 6 In the vehicle type panel the user can select the class of vehicles for which he wants to create a visualization He can select more than one class as this will allow him to make combinations e g if he is interested in all vehicles longer than 5 6m then he has to select gt 5 6m lt 12 2m and gt 12 2m The four buttons at the bottom of the panel allow to respectively select only one class select all deselect all and invert the selection When the select only one class button is pressed then the user can use only one class for his selection If he select a new class then the old one gets deselected This saves some time as a user can quickly jump between subsets without losing his attention to the image Select all allows the selection of all classes of vehicles at once The button selects all but the All vehicles check boxes As it was shown in Chapter 4 the sum of these vehicle groups makes the total sum of all vehicles We decided that selecting all but All vehicles is the better solution than selecting only this check box as it w
44. K make use of it It does not show information about the total number of passing vehicles even though this information is also available INRIX is another product that aggregates data from several different data sources in cluding NDW As a result it can display the current traffic situation congestions It has the same traffic visualization features as Google Live Traffic and Onderweg and also the same problems as described above It seems that there is not much work done and no freely available products that make use of traffic data as number of vehicles that are passing through a street or conjunction This can be either because of the unavailability of such data but also because of the limited applicability of these tools The lack of either the data and the tools makes it very difficult for researchers to understand in detail how the city works 3 http www inrixtraffic com City Traffic Visualization 11 Chapter 4 Data Preprocessing 4 1 Nationale Databank Wegverkeersgegevens NDW De Nationale Databank Wegverkeersgegevens NDW or The National Data Warehouse for Traffic Information in English is an organization that was established in 2007 and involves various authorities working closely together to develop and maintain a traffic data database for the road network of the Netherlands NDW started serving traffic data about the most important roads somewhere in 2009 Since then the databank has grown rapidly and curr
45. SESOR OOE da i 3 Mex SZE A TAS 6 KZ i 2 a A w a a yak i a A je ae a SGI Wee a te ASS p The eko ihoven EP tes ihoven 132 ps ie ar Y A 4 I e e 4 We a A 4 e i A lt p lt gt gt Te et pal R ALZAN 4 e gt A J 328 Ei pasi COD N ve CD 20 Z i Y N2 0 ad r 7 a gt a 7 a R A 4 i a 4 f a J sf 30 0 5500 11000 F 0 5500 7 Lexy 0 5500 a 00 00 03 59 b 04 00 07 59 c 08 00 11 59 Woensel O D 7 el a z A he Sx 0 5500 d 12 00 15 59 Figure 5 16 Traffic intensity for the city of Eindhoven for 24 May 2013 divided into periods of 4 hours Maximum traffic intensity for the color scheme is fixed at 11 000 vehicles As it was mentioned above some of the colors of the schemes might blend with the colors used on the background map It is also possible that a user might want a more clean view with a background map being either less visible or missing at all To solve this we made the background with adjustable transparency This is done by covering the background map with a white screen and changing its transparency and thereby makes the impression as the background map is disappearing User can change this parameter to make the desired image Figure 5 15 Using colors to encode values proves to give a very good overview of the traffic situation but analyzing only the colors makes it practically impossible to determine the exact value
46. Technische Universiteit Eindhoven University of Technology TU Department of Mathematics and Computer Science City Traffic Visualization Master Thesis Valcho Dimitrov Supervisors Prof dr ir Jack van Wijk MSc Mark Stobbe Eindhoven October 2014 Abstract Sooner or later every modern city faces the challenges to cope with increased traffic and street congestions Increased e commerce means an increase in the number of delivery trucks driving around Understanding how a city works can be used by researchers and governments to enforce new policies for regulating the traffic Because of this cities are pioneering with creating measuring sites that collect traffic data for further analysis NDW is relatively young and fast developing system that collects and stores such data for the major roads around the Nederlands Unfortunately there still haven t been many applications that can present the data in easy to understand and analyze format Furthermore the few products that have been released are aimed to be commercial oriented rather than research helping This thesis describes an interactive visualization of NDW traffic data We designed and built a prototype that does has two main functions First it cleans the raw NDW data We remove the unnecessary information and organize the rest in a more useful to us matter This gives a vast reduction in used memory storage space Second we created various types of visualizations
47. To help with this interactive tools which show the data in charts with linear time axis are often used The issue with them is that they can address only one specific case at a time Stephen Few has done lots of work in the field of creating and improving charts 13 14 11 12 Many of his advices are used in this thesis more details are given in Chapter 5 Figure 3 2 Spatial distribution of forest fires for the same period of the year for the span of 25 years Each image aggregates data for the period of 5 years A darker red color indicates larger number of forest fires in that region Images can be arranged in a 2D matrix like shown here or used to create an animation Image taken from 4 In real life scenarios it is often necessary to visualize large data sets Simply showing everything on the map will produce messy and practically unreadable images The main methods used to deal with large data sets include data aggregation and summarization ap plying more sophisticated computational techniques such as data mining and developing projections of the data that move items away from their geographic location to fill graphic space more efficiently Figures 3 2 and 3 3 Apart from the academic work there are several commercial products available around the Internet that aim at visualization of traffic data One of the most popular ones is the Google Live Traffic feature of Google Maps The live traffic u
48. _line table in the database and rename it to tmc_line2 From the SQL editor run the following query CREATE TABLE tmc line AS SELECT gid loc_nr geom ST_AsGeoJson geom FROM tmc_line2 Locate the newly created tmc_line table and rename its columns gid to indez loc_nr to id and geom to coordinates index id coordinates 1 1 2001 type MultiLineString coordinates 4 62176667336509 52 40260236761 2 10 3070 type MultiLineString coordinates 4 5183459437823 51 948429283937 i 100 3990 type MultiLineString coordinates 4 793769419557 52 9130428434369 A P mN FAAA Perm a kha Tea Tame Oe es mar ana i eama TTE 000071 235201 GCG Ca aard DIT 5 Goto http ndw dysi nl and login using your credentials 6 Use the interface to download raw data from NDW and extract files from the archive Download only traffic intensity data Intensiteit for the required locations We rec ommend to download data for 5 minuut aggregation step aggregatietijd For more details about using the user interface once logged in go to Handleiding Historische Exporttool 7 Open City Traffic Visualization and load a connection string file Data Load Con nection String File 64 City Traffic Visualization APPENDIX A ADMINISTRATOR MANUAL 8 Select Data Raw Data Format option from the menu In the new window select all extracted raw data files and press PROCESS Depending on t
49. abel from one or several locations first select the locations Then select the label from the combo box and press Remove Label a VACAS q Selectlabel I NEW LABEL a TJ FL Pa A pe UE i Koliiennowen a Y isenhowerlaan_IM Figure B 9 Creating a label and assigning it to markers Top image label Fisenhowerlaan_IN is being created and is about to be assigned to the selected markers blue Bottom image all markers with label Fisenhowerlaan_IN are shown painted with the color of the label See Label s Markers is similar to the Show Markers option from the Create A Label menu Simply select a label and press Show Markers to see the markers with that label Figure B 10 See Label s Markers option Markers assigned with Fisenhowerlaan_IN label are painted in purple City Traffic Visualization 71 APPENDIX B USER MANUAL Bind Markers To A Road is one of the most important functions in the Map Editor menu It allows the user to bind measuring locations to a road Road 29002 Positive Direction Negative Direction m Positive Negative GEQ01_SRETI043r Lis GEOO1 SRETIO42r IP a g ui E BE aare a Vaal vade EE CEA mit p EK _ rh EF ARC AE Cata EE Figure B 11 Bind Markers To A Road function Markers are assigned to road 29002 Binding a measuring location to a road can be done in two ways The first is semi 12 City Traffic Visualizati
50. an answer questions as 1 Where are vehicles entering the city City Traffic Visualization 3 CHAPTER 2 VISUALIZATION REQUIREMENTS OT R w N Sa 2 Where are vehicles leaving the city Where are vehicles traveling around in the city How many vehicles are going in and out of the city at any time Is there any difference in the number of vehicles and their trajectories between work days and weekends months seasons What is the difference in the behavior of different types of vehicles Are there any exceptions from the normal behavior and can there be identified How does the traffic situation looks like at some time When is all this happening 2 2 Requirements Based on the description of the project we have the following technical requirements R1 R2 R3 The application should take as an input traffic data and produce as an output visual ization images The application should work with historical traffic data provided by Nationale Databank Wegverkeersgegevens NDW The application should provide a view of the data over a geographical map Analyzing the research questions above revealed several more requirements To answer the questions where and when R4 R5 The application should provide temporal and spatial views of the data The application should present the data with respect to time and space Furthermore we can identify also the need to R6 RT R8
51. and index enhancements that apply to these spatial types With its help we can translate Shapefiles into a Structured Query Lan guage SQL file and then import this file into a database After that we can do some more processing to shape the data in the format most useful for us The whole process is as follows Step 1 Translate Shapefiles into SQL file shp2pgsql s lt SRID gt lt shapefile gt lt tablename gt lt gisdatabase gt gt filename sql Step 2 Import the SQL fle into the database psql d lt gisdatabase gt U lt username gt h lt hostname gt p lt port gt f filename sql Step 3 CREATE TABLE lt tablename gt AS SELECT lt column1 columnN gt ST_AsGeoJson lt columnX gt FROM dataTableName If peAdmin III is used to access the database then it is easier to use the graphical user interface to directly perforn Steps 1 and 2 We perform this operation on the RVM network Shape file After the first two steps we have the table presented in Figure 4 5 gid loc_nr symbol geom PK serial integer smallint geometry MultiLineStringZ HM 1 fp f2001 3 01050000C00100000001020000C004000000E1EB23BA7D4913403 2 2002 3 01050000C00100000001020000C004000000E396525A38A60F40E E 2003 3 01050000C00100000001020000C004000000CD247580E2A31140C Figure 4 5 The result table from steps 1 and 2 performed on the RVM network Shapefile Each row stores the available information about one road from the RVM network Columns gid and loc_nr present a uni
52. atabank 31 actuele_verkeersgegevens and download the latest version of the VILD file bottom of the page Bijlagen al NDW Interfacebeschrijving versie 2 2 pdf Erratum NDW Interfacebeschrijving versie 2 2 juli 2013 pdf E Erratum NDW Interfacebeschrijving versie 2 2 januari 2014 pdf ai NDW Interfacebeschrijving versie 2 2 _X5D zip L vo 5 6 a tot 4 juni 2014 zip fl VILD 5 7 a vanaf 4 juni 2014 zip 42 Achtergronden bij NDW data 2014 02 pdf 2 Locate the downloaded file and open it with the help of WinRAR 7 Zip or other archiv ing software Locate tmc line dbf tmc_line prj tmc_line shp and tmc_line shx files and extract them from the archive 3 Import the shape files into the database This can be done through the SQL editor or use the graphical user interface From the pgAdmin III menu select Plugins gt PostGis Shapefile and DBS loader Select tmc_line shp file Make sure that you have entered the correct database and credentials and press Import City Traffic Visualization 63 APPENDIX A ADMINISTRATOR MANUAL ky PostGIS Shapefile Import Export Manager 2s PostGl5 Connection View connection details j PostGls connection Import Import List Shapefile C Users Valcho Desktop tme_line shp Server Host localhost PostGl5 Connection Username postgres Password TITT Database TrafficDatal 4 Locate tmc
53. ation 51 CHAPTER 5 DATA VISUALIZATION 5 4 2 Bar charts data visualization Vehicles per hour Chart type Bars Split None From 2013 11 15 00 00 To 2013 11 1523 59 Agg Step 30 min Vehicles gt 5 6m lt 12 2m gt 12 2m Locations ENTRANCE Chart type Grouping 1000 Bars v Labels vi Vehicle Type Locations El WC Al Vehicles M ENTRANCE 800 M _ lt 5 6m Compare M gt 5 6m lt 12 2m a M gt 12 2m M Uncategorized _ Compare Line 0660 600 Week Days ia a ET ie al i 400 i Aggregation step 30 min i Unit y Scale 3 Total Vehicles m Automatic 200 fa Veh Hour Fixed 860 0 Time Calculation period From 2013 11 15 00 00 H To 2013 11 15 23 59 H 0 Figure 5 29 Bar chart presenting daily traffic intensity for vehicles longer than 5 6 meters at ENTRANCE labeled locations on 15 11 2013 Number of vehicles Chart type Bars Split None From 2013 11 15 00 00 To 2013 11 1523 59 Agg Step 30 min Vehicles gt 5 6m lt 12 2m gt 12 2m Locations ENTRANCE EXIT Compared Compare Line Chart type i Grouping 475 t j Bars RZ Labels v 380 Vehicle Type Locations a M An Vehicles HL ENTRANCE RZ 285 M lt 5 6m y Compare lv gt 5 6m lt 12 2m y n mia M EXIT Ad Wir gt 12 2m i 190 l a M _ Uncategorized x Compare Line 0G60 95 Week Days 0 i i E m ia m_ 95 E Aggregation step
54. base to store the data leads to a scalable solution allowing several users to work simultaneously with the same data City Traffic Visualization 57 CHAPTER 7 CONCLUSIONS 7 1 Future work After the completion of our work there are several issues that remained open for future work Some of them have higher priority than others These are e Increase the performance speed of the application We tried to speed up some of our computations by parallelizing algorithm 3 but we believe that more optimizations can be done in this area The second major issue concerns the quality of the fetch data requests which are sent to the database As we stated the speed depends on the size of the fetched data and the connection speed We believe that moving some of the calculations towards the database by improving the request can speed up the whole process e Deal with incompleteness of the data There are many holes in the data caused by malfunctioning of the measuring stations At the moment we do not apply any data mining techniques to clean it e Apart from cleaning the data there are other clustering algorithms to identify patterns in the data This can be used to find days with similar traffic e We wanted to implement several more visualization techniques mainly connected with comparing data making a calendar view van Wijk 22 Using other type of charts can also provide new insights for the user Finally our curre
55. c 7 Veh Hour O Fixed 0 0 Time Calculation period From 2013 01 03 00 00 E To 2013 01 03 23 59 E Jump period gt Figure B 20 Chart visualizations user interface ol APPENDIX B USER MANUAL B 4 Map Interaction cz a ae SN eee q a sw Location GEOQO1 _SRETIO02r Labels Number Of Lanes 3 Indexes 1C 2C 30 4C 5C 60 TA 130 140 site Name 1 Sterrenlaan de es 9 Site Name 2 gt A ay Kis at AMOS pa EP J adjusted Lat 51 46951675415039 Woe heh lini Adjusted Long 5 487542152404785 Default Lat 51 470001220703125 Default Long 5 4874701499938965 Direction positive JROADNUMBER B 104 ROADNAME John F Kennedylaan FIRST_NAME SI gt a Se z Trzej E el Laas ie 1 AR a Kegs POST R T AA HF A et a i aS Ti T z Controlling the map is fairly easy Scrolling the map around can be done by clicking and holding the right mouse button anywhere on the map and moving the mouse while still holding the button Releasing the button stops the scrolling Adjusting the zoom level can be done in two ways First Zoom level controllers can be used They are located at the bottom left corner of the map Dragging the slider up or pressing the button increase the zoom level map comes closer and dragging it down or pressing the button decreases the zoom level map gets further away Optionally the user can use mouse wheel to adjust the zooming of the map
56. ch day These lines are also made thicker in order to be more noticeable On figure 5 26 Saturday traffic is shown in yellow and the average traffic in dark yellow We distinguish two cases in splitting data based on location In the first one each measuring location is used to create a line for the chart In the second case we use aggregated data from several locations assigned with the same label Each group label of locations is used for one line on the chart When we split on measuring location then each line is assigned color from scheme c from figure 5 24 If there are more locations selected than colors in the scheme then the colors are reused over and over again Figure 5 27a has twelve lines while the scheme has only nine colors In this case red blue and green are used for two lines If the user splits on label then the color assigned to each label is used Assigning a color to a label is done in the creation of the city model It is important for the user to select colors such that lines 48 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Number of vehicles Chart type Lines Split Vehicle From 2013 01 03 00 00 To 2013 01 03 23 59 AggStep 30min Vehicles All Vehicles lt 5 6m gt 5 6m lt 122m gt 12 2m Uncategorized Chart type Grouping Locations ENTRANCE 7500 Lines X Labels x All s
57. data files has to be manually uploaded to the database The visualization part of our application allows the researcher to use the techniques from Chapter 5 to generate clear and meaningful images Researchers can inspect these pictures and change visualization settings to offer more insight 56 City Traffic Visualization Chapter r Conclusions In this thesis we presented our interactive prototype for visualizing NDW historical data Proper visualization images can be used by researchers to understand and improve the traffic of the cities in the Netherlands Below we present our conclusions and some recommendations for future work and improving our work We started this project with a few general requirements and several research questions that had to be answered We tried to specify our goals and address each challenge Every design decision which has been taken is done to make our application meet with the requirements Working with big data always brings some limitations to the developed products The main problems come from limitation in the memory storage capacity performance speed and limitations of visualization techniques NDW provide data with lots of duplicating records As shown in chapter 4 the useful and unique information of the raw data downloaded from the databank can be less than 6 We used some simple techniques to remove the extra data and summarize the remainder into tables Our approach makes it easier to access
58. data from two difierent groups of measuring points Figure 5 30 The main advantage over our line charts is that it draws a line show ing the difference between the corresponding values This is for instance useful to compare incoming and outgoing traffic 5 4 3 Charts visualization user interface The user interface for charts visualization is to a great extent similar to the one used for map visualizations Figure 5 32 This can help the user to adapt faster and make it easier to use the application Panels C D F G and H have the same functionality as in the map visualizations user panel There are a few differences though The small split radio buttons allow the user to select where he wants to make the data split The two versions of panel D are used respectively for line charts and bar charts The bar charts panel has also controllers which allow the selection of the compare mode Panel E allows the selection of week days It is active only if the user select to make split there and in line charts visualization Both panels C and E has small colored buttons in front of every check box The colors of the button represents the color binded to that option Showing the colors there can help the user interface panel act as a sort of legend for the chart The two top panels A and B allow the selection of chart type lines or bar and grouping of the data either using labeled locations or individual locations The aggregation step allows the
59. dents and incidents on all the national roads City Traffic Visualization 13 CHAPTER 4 DATA PREPROCESSING e Safety related announcements such as a wrong way driver that are issued by the traffic control centers e The status open closed of bridges e The status open closed of peak and regular lanes The database is primarily intended for the NDW partners themselves Other parties that want to access the historical data need to acquire a license which can happen after submitting a request to NDW for this 4 2 Traffic Data This master thesis was done using the data that was available with the university license This allows us to access only the historical traffic information and hence the developed application is making use only of this data As we mentioned above the data is collected on the specific measuring locations of the road network Hence the data which is available for download is grouped by measuring locations NDW has a very well organized user interface which allows you to request only the data which is of interest to the user This can later save you processing time and storage space for the data itself The most important selection filters which can be used by the user are Select the locations for which you need information Select the time period for which you need the information Select which type of information you need intensity of the traffic average speed of the vehicles and average travel time
60. e are mapped evenly to the values between 0 and maz value We conducted a small research amongst members and researchers from Smart Logistics lab as our main target group and also some students from other work groups We showed them images similar to figures 5 12 and 5 13 asking them for their opinion about 1 if the color scheme is useful 2 if you see the difference between colors traffic intensity easily 3 which scheme is the best one We did not get a clear winner about the best one although Cyan Magenta and Yellow Green Blue were least preferred Two groups of users were identified one that recognizes the more colorful schemes as being more useful and one that finds less colorful schemes more useful Subjects from the first group even pointed Blue Green Red as being the best scheme despite its flaws It seems that different users have different preferences and different ways of reading colors Because of this reason and because all ColorBrewer schemes used are designed to be more printer and photocopy friendly none of the proposed color schemes was dropped and all of them were implemented Discrete versions of each color scheme are also available Discrete schemes provide a smaller number of colors for drawing the roads and hence streets with similar intensity are colored the same way This can help in the traffic analysis by making it easier to identify sections with close traffic Figure 5 14 In our tool discrete color schemes have
61. e d 10 if vehiclesPerHour then 11 result sum total Lines 12 else 13 result sum totalLines x 1 60 14 Save result in l The pseudo code presents shortly the algorithm which we use for calculating the final result for each location The inputs for the algorithm are list of measuring locations L list of the selected vehicle classes V the selected time period T and boolean flag vehiclesPerHour The location list L contains only the locations that will be used for the visualization e g locations that are selected locations linked to roads This helps us to decrease the total computation time The time period T and the list of vehicle classes V is used to reduce the size of the data that is need for the calculations The data is obtained from the locations l http www openstreetmap org City Traffic Visualization 29 CHAPTER 5 DATA VISUALIZATION tables described in Chapter 4 The boolean flag vehiclesPerHour shows the form of the final result The result can be given in either vehicles per hour or total number of vehicles According to NDW vehicles per hour veh h is the standard measurement unit for traffic intensity Because of that the data provided is given in this unit Hence in our locations data tables we also store the traffic intensity in veh h If the user selects this option then he gets the aggregated average intensity line 10 in the algorithm 1 1 lope 5 1 Where Iaggr is the avera
62. e roads or close to the location of the point at which they measure traffic Use Bind Markers To A Road to allocate measuring locations to a road Make sure that you have allocated each location to the right direction of the road Use the metadata for each point to determine its correct location The traffic on each road is determined by the traffic from the measuring locations linked to it This step should be done carefully and with utmost attention If needed use Edit Marker Data to correct the metadata associated with the measuring location Use Create A Label to assign labels to the measuring points This allows you to visualize traffic from a group of points Save your model for future use by selecting Save Model from File menu City Traffic Visualization 15 APPENDIX B USER MANUAL B 3 Create a view Views allow you to get a visualization of the selected information Both types of visualization have a similar user interface Before starting make sure that you have loaded a connection string file Secondly create a new city model or load an existing one Map visualizations present traffic data over geographical road map ag EN om 7 T p jt i 5 J A i i Ra 38779 d Woensel Moord i 39362 pA ji Eur 36889 Tenitaen srerreniasn Woensel zug 7 35396 ae 35316 j a 34763 151831 7 TRE eT AE a W STE 77 Eindhoven kal iry E 1 q A A 7413455
63. e the number of these locations will increase In consultations with members from the OPAC group we agreed that until this happens they cannot make good use of the non intensity data City Traffic Visualization 19 CHAPTER 4 DATA PREPROCESSING 4 3 Roads Network Data Road authorities have selected pathways that are important for the accessibility of the region They called that network RVM network Regionaal VerkeersManagement Regional Traffic Management NDW is using this network to place measurement equipment and collect traffic data This information is useful to us and can be used it to visualize the traffic data sosna LU RYB REST KA ack Ne CA AL ZEW U SORE CAE jr YA A BE gt gt LONA OE ER ION SNe SAR DATA NJ N A i Suy gt RR AP ERAN AX Da Shes TA OS ET ar Sai TK OS OIA Figure 4 4 RVM network from June 2014 The RVM network is available for download on the NDW website in the form of a Shapefile shp For our application we have to change the format of the file to make it easier to read and edit This was achieved with the help of a PostgreSQL database with PostGIS extension 20 City Traffic Visualization CHAPTER 4 DATA PREPROCESSING More technical details about the database used are available later in this thesis Chapter 6 and Appendix A PostGIS adds extra types geometry geography raster and others to the PostgreSQL database It also adds functions operators
64. eduling of ICT infrastructure people and governmental policymaking 21 Smart Logistics equals 3P I i e Planning People Policy and Infrastructure and is the synchronized interplay of these four key domains The Transportation Chain consists of three parts the Pickup Chain or the first mile the Transit Chain and the Delivery Chain or the last mile The Smart Logistics Labs research is completely embedded in the Transportation Chain working on topics related to separate parts of the chain and on the integration of the separate parts The challenges in the Transportation Chain are to efficiently handle increased uncertainty and high complexity This is important since the real life world does not fit into a deterministic and static straitjacket which is assumed by many published models and industry tools Any decision action plan or schedule built on unrealistic assumptions is bound to be less than optimal once realized Understanding how the city works in reality can be a key factor in producing better and better solutions to transportation problems The road network of the city is its arteries It is important to know its condition at any time in order to maintain it optimize and expand it and use its full potential For that it is important for a researcher to be able to answer the following questions 2 1 Objectives In this thesis we try to visualize NDW traffic data and develop an application with the help of which users c
65. ements of each list in wLists n 19 Add avg information to list avgList 20 Add avgList to vLists Algorithm 3 calculates one list of values for each selected day in the selected interval It uses algorithm 2 for some intermediate calculations line 10 Flag avg indicate user s preference if he wants to see also the average values for each group of days e g average traffic for all Mondays in a month lines 16 20 Flag avgOnly indicates if the user needs only the average values To visualize our calculated lists of values on a line chart we used many of the advices given by Stephen Few in his papers and book see Bibliography We paid particular attention to chart size line size lines on the plot axis lines background lines and color number of labels on X and Y axis and legend Some of the visualized values have a small label below them and below the x Axis showing the time stamp of the visualized value For convenience we try to put labels for as many values as possible Having in mind that there is only limited number of labels that can be located along the xAxis without overlapping we use formula 5 5 We put a label under the first visualized value and then we check for the rest of the values If there is enough room for a new label we put one This can result in our last point to be left without a label We believe that this will not be a big issue in reading the graph as the user knows that the last value is always at the end of the
66. ently it ields more than 200 T B of data from 20 400 measurement sites spread over 6 000 kilometers of roads Some of the main priorities are insuring high quality of the data and uniformity NDW ensures that all traffic data is entered properly into the historical database If there are any errors or delays in the real time traffic data they are being corrected before entered in the database Uniformity is achieved by presetting some rules and agreements about definitions and calculation methods This insures that the data coming from the different partners has the same shape and is easy to use We have to note that it does not mean that we have 100 error free and available data NDW provides two types of information traffic information and status information Each measuring site collects data every minute and then made available to the users in around 75 seconds Traffic information includes e Traffic flow the number of vehicles that pass a measurement site within a certain period of time e Average speed e Realized or estimated travel time e Vehicle class derived from the length of the passing vehicles All the accumulated real time data is saved in the historic record and is thus is available for later use and consultation The status information refers to the availability of a certain road e Road works and event related traffic measures on virtually all the roads in the Nether lands e Reports of congestion acci
67. es and locations either measuring locations or labels and they are clearly defined Making a split on the time dimension proved to be more delicate Data can be divided into hourly data daily data monthly etc Based on our consultations of Smart Logistics lab researchers and based on the requirements which we have derived we concluded that it is important to make a split based on days and also to be able to identify which day of the week corresponds to each data stream Hence in our tool we focused on improving the data visualization with a daily time split Splitting on longer periods of several or more days is left for future work Despite that we believe that with the current implementation we are still able to give an answer to almost all of the asked questions Each visualization image is 44 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION generated in the following steps 1 Calculate lists of values that are to be visualized 2 Find the location of each value on the chart plot 3 Connect consecutive values from each list with a colored line As with the data visualizations on a map here we also start with calculating the values that are to be visualized The algorithms used for that are very similar to algorithm 1 and are modified versions of it They differ depending on the selected split dimension The user is allowed to make a split based on vehicle class location day of the week or both vehicle class and day of
68. face Version C shows up when we have bar charts and show data for labels Combo boxes 8 and 10 allows the selection of labels for the visualized data If check box 9 is selected the user can compare the data from the two selected labels figure B 19 If check box 11 is selected then an orange compare line is drawn showing the difference between the height of the corresponding bars Positioning the mouse cursor over a line or a bar highlights it and a small box is shows containing the visualized value at that point Number of vehicles Chart type Bars Split None From 2013 11 15 00 00 To 2013 11 1523 59 Agg Step 30 min Vehicles gt 5 6m lt 122m gt 122m Locations ENTRANCE EXIT Compared Compare Line 380 190 190 380 Figure B 19 Traffic data visualization with the help of bar chart 80 City Traffic Visualization 2a 2b 25555 de 2d City Traffic Visualization APPENDIX B USER MANUAL R SEE m Locations Chart type Grouping 3 Split B Lines v labels w GEO01_SRETI042 EG01_S5R All split 2c GE001_SR Vehicle Type Locations GEO01_SR E001_SRETI043 i GEOO1 SRETMT2 a a CZELE Onewa e SRF Locations Uncategorized C Split C 2 8 A Week Days 3 _ v Compare se 1 mn 11 _ _ Compare Line IM Li Mod mst Mil Sa 111 Average ab Tuse L Fr 5g L Su 12L only Avg i Unit y Scale a Total Vehicles Automati
69. from PoiintsMetaData Table 4 2 Roads are drawn using the coordinates from tmc_line Figure 4 6 as a guideline and simply connecting them with a line Figure 5 2 For automating the process we calculated the distance between each measuring location and each road in our model next we linked the location to the road it was closest to As Figure 5 2 explains this resulted in linking some locations to the wrong road The problem occurred because some locations were positioned closer to one road although they were mea suring the traffic on another road To solve this problem the user has to first identify the problematic areas and then either redraw some of the roads so that they go further from some locations or change the drawn position of the points so that they are closer to the road they belong to Unfortunately this would take too much time and efforts for the user to correct everything and it is also likely that he might miss some of the errors Because of that we decided to drop this solution and let the user manually assign location to road Although this can also require a lot of effort it also has the advantage that each linking is inspected by the user upon building the model Another advantage is that the user can choose which locations and roads he wants to link together The unused objects will not be viualized hence reduce clutter in the image Figure 5 3 gives an example with a few measuring locations Each measuring location
70. from industry interested in historical traffic data but they should own at least the skills already mentioned Users select use inspect k Te gt Crn Figure 2 1 Basic workflow for City Traffic Visualization application City Traffic Visualization 5 Chapter 3 Related Work We are interested in showing traffic data taking both its geospatial and temporal proper ties into account Route maps are one of the most common forms of presenting geospatial data Cartography has been around since the dawn of human civilization In the past several decades lots of work has been done to improve our maps and to be able to fit more and various information Unfortunately most of the work that has been done in the field of visualizing traffic data on a map is in the case where the data itself consists of object trajectories Never theless some of the visualization techniques and recommendations that have been developed can be employed in other projects Visualizing the spatial component of the data is usually done by presenting the information on the map on the locations where the event for which we have measurements is happening This can include drawing the route of a vehicle or showing the average vehicle speed at the point where a speed camera is located Most traditional for the visualization of traffic movement are arrows or flow lines drawn on a map or image Figure 3 1 Lr Pith i e CT KI Pu ZMR
71. ge intensity for the whole period N is the number of measurements used and J is the intensity for measurement i Our target group of users is mainly interested in the total number of vehicles that are passing through some point Because of that we added this option for the final calculations result line 12 3 I N 60 Where Viota 18 the total number of vehicles for the whole period Nis the number of mea surements used J Intensity for measurement i and T is the length of the time period in minutes Calculated results are stored for each measuring location in order to be used for the visualization Viotal 5 2 Unit m Total Vehicles Ci Veh Hour Figure 5 7 Unit panel used for selecting the measurement unit for the final result 5 3 1 Visualization of traffic over the road network The first map view which we implemented is a traffic view that colors the road network depending on the traffic intensity Figure 5 8 The visualization image is generated in four steps 1 Calculate the values for each location 2 Calculate parallel lines on both sides of each road which are used to display the traffic in each direction 3 Draw the roads using colors respective to the visualized values 4 Draw locations values and arrows if selected Step 1 is explained before As shown in Figure 5 8 we use two parallel polylines on a distance d on each side of every road to present the traffic in each direction To calculate
72. he size of the raw data this can take from several minutes to a few hours Open Files v Ew_2013_intensiteit 00179 csv Ehv_2013_intensiteit_00180 c5w 9 After the import is complete check your database If the work was done successfully the database should now have many tables one named after each measuring location one table named pointsmetadata and one table tmc_line Depending on their privileges users can also use City Traffic Visualization to perform steps 6 to 8 and import raw data into the database City Traffic Visualization 65 Appendix B User Manual P City Traffic Visualization lt H Welcome to City Traffic Visualization Explore traffic data in 3 simple steps 1 Load a connection string file and import raw data into the database 2 Create a new model or load an existing one 3 Use the Views to explore the data Connection string file D Google Drive TUe_GRADUATION Projects City Traffic Visualization Connection String txt Figure B 1 A Menu B Working panels C Connection string path In the main application window frame three areas can be recognized In top is the menu A in the middle are situated the working plot panels B where every opened view is put on the bottom the path and the name of the currently loaded connection string file are shown C 66 City Traffic Visualization APPENDIX B USER MANUAL B 1 Menu B 1 1 File Menu File Data Map Editor Wi Create New M
73. ill allow the user to make faster selection of at least three vehicle classes together e g if we are interested in all vehicle but the unknown then we press select all button and then deselect the one which we do not need Deselect all allows us to quickly clear our selections The invert selection button helps to quickly invert a selection e g unselected classes get selected and selected classes get deselected The locations panel helps the user to select the locations from which he wants to obtain data It allows the selection of either individual locations or group of locations having the same label The time controllers panel is inspired by the work of Andrienko 6 They propose the idea to add buttons bottom of the panel that enable the user to quickly jump forward and backward in time This can be used to simulate an animation and allows spotting time variances in the data The exact jumping intervals are explained in the next sections Date 28 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION selectors have a button next to them which opens a calendar This allows quick and accurate selection of a date These three groups of controllers compose the main part of our user interface The other part of the interface varies depend on the type of visualization chosen by the user The new parts are explained in next sections of this chapter The three main groups of controllers also have minor variations depending on
74. in segment AB is part of the polyline paz Figure 5 10 Cluttering triangle at the curve of the road caused by the order of adding points to the polyline points list The sharper the road curve is the bigger the triangle is The numbers indicate the order in which the inner polyline is drawn For each location we find its projection onto the polyline The segment of the polyline located between two projections R and B Figure 5 11 is drawn in the color determined by location R assuming that we draw the polyline in direction from R to B One of our aims is to draw the user s attention to areas with heavy traffic Therefore we also change the thickness of the line segment depending on the traffic there Roads with heavier traffic are drawn with thicker lines and roads with lighter traffic with thinner For drawing each road a polyline with 1 pixel thickness is used It is made as thin as possible because we don t want it to distract the user The thickness of colored polylines can vary from 5 pixels to 10 pixels Even the thinnest segments are significantly thicker than the road because we want them to be big enough in order for the user to be able to determine their color The maximal thickness was chosen such that it is enough to create a readable image but in the same time small enough so that the roads on the map are still visible The color schemes for the roads are designed to be easy to understand and to draw the attent
75. ines ColorBrewer also gives examples of several qualitative color schemes Based on these recommendations we designed three color schemes that we use depending on where the user makes a data split Figure 5 24 Colors City Traffic Visualization 4T CHAPTER 5 DATA VISUALIZATION Algorithm 4 Calculating y Axis labels Input Number of labels nrLabels min Value max Value Output List yLabels max range maxV alue minV alue unrTickS 5 x logio unrT ickS 2 powl0x 10 rTickS Pouce x powl0ax 5 max nr Labels 1 rTickS for i 0 nrLabels 1 do label minValue i rTickS Add label to yLabels a Split on vehicle class b Split on day of the week c Split on location N O OE W M HH ua Figure 5 24 Default color schemes that are used for coloring the lines in our line chart from scheme a are assigned to each vehicle class and the colors from scheme b to each day of the week If a split is made based on vehicle class then lines are drawn using the colors assigned to each class Figure 5 25 if a split is made based on days of the week then the colors assigned to the days are used Figure 5 26 The user can also select to see the lines presenting average traffic data for each week day e g average traffic intensity for Mondays The lines that show the average traffic are drawn using a darker version of the same colors used for drawing ea
76. ion to the important areas The main questions which are to be answered with the help of our tool involve identifying heavy traffic Hence we need color schemes which high value colors are easily distinguishable and more noticeable We tried to solve one of the problems which we mentioned about Google Maps color scheme light colors are located somewhere in the middle of the scheme Here they are put at the beginning so that the higher the value 32 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION Figure 5 11 Drawing polyline P in different colors and thickness The polyline has two measuring locations associated with it R and B First the projections R and B of these locations onto the line are found The arrow indicates the direction of drawing of the polyline The polyline is drawn from the beginning using the default color and thickness After each projection the parameters are changed depending on the value of the respective measuring location The colors and thickness used in this figure are for demonstrative purposes only In our application location points are drawn in the same default color the darker the color gets The high end colors should also be clearly distinguishable from the colors used in background map We also want color schemes to make a transition between at most three different leading colors The small number of colors makes the visualization easier to understand and more clear Furthermore the human eye
77. is project Following the requirements we had to improve the old solutions and develop some new ideas and techniques that make a better use of NDW data and suit Smart Logistics researchers needs Following the requirements for temporal and spatial views R3 4 5 7 9 and 10 we designed two types of views that we believe to satisfy of them The first one visualizes data over a geographical map The second one visualizes data as time series depending on the selected data and time periods Both views produce different type of images but result from projecting the same city model and controllers for filtering the data 5 1 City Model The city model consists of roads measuring locations and some extra meta information which is added by the user and can be used later for further exploring The model aims to aid the visualization process by first constraining the number of data sources that are used take and process data only from locations that we want to explore reduce the size of the road network which we want to visualize and combining measuring locations into groups In this section we explain the model in more detail More description about our tool functionality and detailed description how to create a city model can be found in Appendix B Figure 5 1 shows the difference between all the available data for the measuring locations and RVM network and the trimmed data which is used to create a model The created model helps us to focus
78. length as AB and are on a distance d from AB The newly calculated points A and B are added to the list of points that describes the polyline used for traffic in positive direction and A and B respectively to the list of points that describes the polyline used for traffic in negative direction assuming that AB is the positive direction This can cause some clutter problems as we can see on Figure 5 10 As the road makes a curve the polyline which is on the inner side of the curve makes a small triangle The sharper the curve is the more obvious this triangle is We can see that the triangles are relatively small compared to the length of the roads and in practice these are hardly noticable Figure 5 8 Nevertheless a better soutions should be implemented The next step in generating the image is to color the new parallel polylines to show the values calculated for the measuring locations In the construction of the city model we linked each location to a road and indicated in which direction that location is recording the traffic Each of the two parallel polylines has a direction To draw a colorful polyline we need the locations that measure traffic on that road and in the same direction indicated by the polyline City Traffic Visualization 31 CHAPTER 5 DATA VISUALIZATION Figure 5 9 Drawing parallel lines A B and A B lines on both sides of segment AB at distance d from it and with the same length The ma
79. lrc A Ej nd hoven omakkers Wes Strijpse 5 MA vs E 1 Vondertk rier Na TN A ip gt 7 Es 3 Ru dze Engelsbergen zergen naali D 148724 JP Zuid isbuurt 11901 116980 Hagen 5393 Rochusbuurt ide Spoorbaan z a Elzent Noord k SIOWMIWersbuurt Tuindorp 1122 16156 gt Elzent Zuid 14606lorisk want es j enderd Lonakker pi 1 ji Severijnpark p menplein H Jieuwe Erven Gerardusplein Kruidenbuurt ievoet N m po 1 ms thy We tops i 0 13184 26368 Figure B 16 Traffic data visualization over measuring locations Group 4 of check boxes allows the selection of vehicle class Vehicles are divided into classes by their length lt 5 6m 4b gt 5 6m lt 12 2m 4c gt 12 2m 4d and Uncategorized 4e By default you can select several boxes at the same time Pressing button 5 allows the selection of only one class Pressing button 6 selects all vehicle classes 4b to 4e Button 7 deselects all selected classes and button 8 inverts the selection selected classes get deselected and vice versa Table 9 keeps a list of all measuring locations in the model During measuring locations visualization the user can select for which locations he wants to see data Using combo box 10 the user can select locations having the same label Controllers 11 to 13 allow selection of extra objects on the map Box 11 shows the yellow City Traffic Visualization 7
80. n each new table store all measurement results for this location Each table is assigned the name of the location and has the columns shown in Table 4 3 column title periodStart y Index N Table 4 3 List of the column names that can be found in each location data table If we turn our attention back to Figure 4 1 and column Q which shows the intensity of the traffic for the measuring period we will notice a strange anomaly The first four rows of the table contain all measurements made at GEO01_SRETIO43r location The first three rows show the number of vehicles with length lt 5 6 meters lt 5 6 lt 12 2 meters and gt 12 2 meters The last row contains the total number of vehicles that have passed through that location The common rule is 4 1 18 City Traffic Visualization CHAPTER 4 DATA PREPROCESSING N vehicles lt 5 6 N vehicles lt 5 6 lt 12 2 N vehicles gt 12 2 N All vehicles 4 1 In our case for the number of passing vehicles we have 15 0 0 and 60 but 15 0 0 60 This anomaly happens quite often in the data After sending a question to NDW we got the following explanation Vehicles for which no length could be determined because they were too long or because they were changing lanes over the loop or any number of reasons are counted for anyvehicle but not for any vehicleclass Hence the extra vehicles that are counted should belong to the Uncategorized group In ou
81. ncy gt H 8 Bzz Li Figure 5 15 Traffic intensity visualization using the Cyan Magenta color scheme and four levels of transparency for the background map 36 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION connected with a higher value Some of the shades in the middle of the scheme can be con sidered similar to the background map colors used for highways light purple and buildings shades of gray This issue can be avoided by adjusting the brightness of the background map explained later in this chapter Nevertheless high value colors cannot be mistaken and clearly identify high traffic intensity Figure 5 13g The last color scheme implemented is Blue Green Red Figure 5 12h Our initial idea was that since it makes a transition between all main colors of the RGB model it will be easy to see the difference between the remote color shades If we check the image on figure 5 12h and pay more attention to the colors which are slightly left and slightly right from the middle we can conclude that they are very similar although there is a significant distance between them This can mislead the user while analyzing the image According to us this is the least useful scheme Mapping values to a color happens in the following way First we find the highest cal culated value Then it is mapped to the last color in the color scheme and zero is mapped to the list color The rest of the colors in the color schem
82. nd in our cities Governments are forced to create new policies for regulating the increase in traffic keeping our cities functional and safe Improvements to infrastructure and better city planning also play a major role in smart logistics Modern cities have many measuring sites for collecting real time traffic data Every minute they collect data about traffic flow realized travel time estimated travel time traffic speed and vehicle classes For smart logistics it is important to understand how the city works What are the main problems within the city What is limiting an increase in traffic What is the effect of new policies such as environmental areas delivery time windows etc The huge amount of data collected every minute poses a challenge to the experts as it becomes more and more difficult to analyze it and extract the useful information Meanwhile due to the increasing number of measuring sites and upgraded hardware the collected data becomes more complex to be analyzed Nationale Databank Wegverkeersgegevens NDW collects such data summarizes it and presents it to researchers to investigate Unfortunately it is relatively young service and its limited user group slows down the development of software applications which make use of the data Furthermore there is no universal method to read and understand the traffic as different users are interested in different properties of the data In this project we aim to provide several vi
83. ns in the city of Eindhoven for May 2013 and the second one with the same type of data but for the whole 2013 The task was to generate the image on Figure 5 8 Using the file storage system with monthly data calculating the image was done in between 3 and 4 seconds With a yearly data image was done in between 26 and 30 seconds Running the same test using a database located on the testing machine resulted in presenting the image in around 3 seconds regardless of the raw data size We set up a database on a remote server and this slowed down the system as the visualization was done in 12 seconds no matter the size of the database In this case the performance speed depends on the size of the requested data and the connection speed between the database server and the user s work station Hence using a database as storage system does not always result in a major speed up Despite this we kept this solution because of another advantage Once a database server is created it can be used by many users simultaneously each one having his own workstation The application is developed using Java and has two main functions one is to process raw data downloaded from NDW and store it in the database and the other one is to visualize the stored data using various techniques The processing part chapter 4 is intended to be used by the system administrator or any other user who has credentials to write in the database As mentioned before some of the
84. nt visualization approach can be improved by providing the user more interaction and options to better tune the data selection Moreover implementing a search option would help users to find faster measuring locations or roads based on the search criteria and the meta data they provide location name street name number of lanes etc 58 City Traffic Visualization Bibliography 1 Onderweg http www onderweg n1 2 Aigner W Miksch S Schumann H Tominski C Visualization of Time Oriented Data Springer London Dordrecht Heidelberg New York 2011 3 Andrenko G Andrenko N Interactive Maps for Visual Data Exploration Interna tional Journal of Geographical Information Science pp 355 374 1999 4 Andrenko G Andrenko N Demsar U Dransch D Dykes J Fabrikant S Jern M Kraak M Schumann H Tominski C Space Time and Visual Analytics International Journal of Geographical Information Science pp 1577 1600 2010 8 5 Andrenko G Andrenko N Wrobel S Visual Analytics Tools for Analysis of Move ment Data ACM SIGKDD Explorations Newsletter Special issue on visual analytics pp 38 44 December 2010 7 8 6 Andrenko N Andrenko G Gatalsky P Supporting Visual Exploration of Object Movement In Proceedings of the working conference on Advanced visual interfaces AVI 00 pp 217 220 2000 28 7 Andrenko N Andrenko G Voss H Bernardo F Hipolito J Kretchmer
85. o the database Be careful with any changes as they are irreversible Pressing the CANCEL button discards any changes that have not been saved and closes the menu 68 City Traffic Visualization APPENDIX B USER MANUAL Location GEO01_SRETI66r Number of Lanes 1 co Indexes 18 2B 38 4B 1A Site Name 1 Frankrijkstraat Site Name 2 Default Lat 51 45309829711914 Default Long 5 46304988861084 positive B105 Boschdijk Kronehoefstraat Figure B 6 Meta data editor Selected location s data blue point is shown in the text fields on the right where it can be altered The Draw Roads function allows the user to add a road to the road network Clicking with the left mouse button on the map sets a start point for the road Every left mouse button click after that adds the location of the mouse cursor to the points which the road is going through The mouse cursor is always connected to the last set point of the road To remove a point click with the middle scroll mouse button anywhere on the map This removes the last point of the road To finish drawing the road click anywhere with the right mouse button Figure B 7 Draw Roads function The Edit Roads function allows the user to cut and combine roads To cut a road into two pieces first select the Cut option from the menu on the right Then click on a road to cut City Traffic Visualization 69 APPENDIX B USER MANUAL it at the loca
86. odel Alt Open Model Save Current Model Alt 5 Export Image Exit Figure B 2 File menu 1 Create New Model Start building a new model Open Model Open an already existing model Save Current Model Saves the current model Export Image Saves a picture of the current view to the computer no e w N Exit Closes the application B 1 2 Data Menu Map Editor Wiews Help Raw Data Format ANA Load Connection String File Alt L Figure B 3 Data menu 1 Raw Data Format Process and upload raw data to the database 2 Load Connection String File Selects a connection string file for use B 1 3 Map Editor Menu File Data Map Editor Help ella Ms Figure B 4 Map Editor menu Buttons from left to right move marker edit marker data draw roads edit roads create a label see label s markers bind markers to a road delete The options in this menu are used to create a new city model and edit an existing one if needed Steps how to create a model can be found in section B 2 City Traffic Visualization 67 APPENDIX B USER MANUAL 1 Move Marker button allows adjusting the display position of a measuring location marker 2 Edit Marker Data button allows editing the meta data for each measuring location 3 Draw Roads button allows custom drawing of roads on the map pa Edit Roads button allows editing of a road by either cutting it or combining it with another Create A Label bu
87. om cyan to magenta The reason to pick only two colors is that in this way we have smooth transition between the neighboring shades without sudden jumps Cyan and magenta are two of the main colors in the CMYK color model This makes them easily distinguishable from each other Another reason for exactly these colors is that they are not used in the background map and that they attract user attention more easily We choose our scheme to start with cyan because the magenta color is darker and hence more intuitively http colorbrewer2 org City Traffic Visualization 33 CHAPTER 5 DATA VISUALIZATION 34 a Partial Hot Body b Yellow Orange Red c Reds d Yellow Green Blue e Red Purple Purple Red g Cyan Magenta h Blue Green Red Figure 5 12 Color schemes that are used in the tool City Traffic Visualization CHAPTER 5 DATA VISUALIZATION i p o AF Veldho 4 AY kosi a Y NE a Partial Hot Body 22 0 Wo g a En Blue acen Red Figure 5 13 Visualizing the same data as on Figure 5 8 using different color schemes City Traffic Visualization 35 CHAPTER 5 DATA VISUALIZATION Veldha 1968 DS zma a Continuous b Discrete Figure 5 14 Traffic intensity visualization using continuous and discrete version of the Yellow Orange Red color scheme Z rm in O 29 HM 1 u TD 111 c 60 Transparency d 100 Transpare
88. on APPENDIX B USER MANUAL automatic and the second one is fully manual For the semi automatic way first select the road you want to assign markers locations to The selected road is painted in red and its name is shown in the Road text box on the right Second select all markers that are to be binded to this road Press the Auto button to make the application automatically determine the direction of the traffic at the marker and bind the marker for the right direction of the road Markers assigned to the positive direction of the road are shown in the Positive Direction table and the ones in negative direction to Negative Direction Positive direction is the direction in which the road is drawn In almost all of the cases this direction matches the direction in Direction field of the meta data for the marker See figure B 6 Pressing the Auto button just makes the application compare the directions After assigning has been finished carefully inspect the result Make sure that the locations are binded to the right direction Pressing Positive and Negative radio buttons makes small red arrows appear on the road indicating respectively positive or negative direction and markers from that table are painted in blue If a marker needs to be transfered from one direction table to the other simply select this marker and press gt gt or lt lt buttons Assigning locations manually is done in a
89. on step which is used to aggregate our data The algorithms that are used when the user wants to have a split based on location or both vehicle class and location are similar to algorithm 2 The main difference is that when split on location is needed then we sum the data for each of the selected vehicle classes together and if split on both criteria is needed then we do not sum it City Traffic Visualization 45 CHAPTER 5 DATA VISUALIZATION Algorithm 3 Pseudo code for generating values lists making split on days of the week 1 Input List of locations L List of vehicle classes V Time period T tpegin tend 2 List of week days W aggregation step tAgg and flag vehiclesPerHour avg avgOnly 3 Output List of list of values vLists that are to be visualized 4 List lt List lt Value gt gt vLists 5 for week day win W do 6 louf tbegin T List lt List lt Value gt gt wLists 8 while tuf lt teng do 9 if teuf getDay w then tour tour 1day continue 10 List lt List lt Value gt gt buf List Algorithm 2 L V T tAgg vehicles Per Hour 11 List lt Value gt resultList X corresponding elements of each list in bufList 12 Add day of week information to list resultList 13 Add resultList to wLists 14 touf tbuf lday 15 if avgOnly false then Add all lists from wLists to vLists 16 if Avg true then Fe n number of elements in wLists 18 List lt Value gt avgList X corresponding el
90. one by enabling the user to select a group of points and apply a label to them It is practically possible that one or more locations fit in more than one group e g Traffic nf Boschdijk street and Traffic entering the RING thus we allow applying several labels to a measuring point The user can also elects a color to represent each label On Figure 5 5 two labels are applied to the measuring locations Label ENTRANCE green groups the locations that measure the incoming to the city traffic and label EXIT red the locations measuring outgoing traffic We tried to make the selection carefully and identify the correct label for each location based on our knowledge of the city We believe that the quality of our selection is sufficient and we use the specified labels in some of our demonstrations later in this thesis City Traffic Visualization 25 CHAPTER 5 DATA VISUALIZATION Figure 5 3 Example of three measuring loca tions yellow green and red Each location consist of several measuring points each mea suring point collecting information for one line of the road Each measuring location is com bining traffic information about the same type of lanes Yellow location four lane road all lanes are provided for traffic going forward Green location four lane road the two lines are provided for traffic going forward Red lo cation four lane road the two lanes are for traffic leaving on the right side Image f
91. one interface saving time and efforts The interface on figure 5 21 is designed to be organized simple clean and reusable All controllers which have a similar type function or are relevant for making a selection are grouped in panels shown with a red rectangle and capital letter Panels C D and H have already been described in section 5 2 One can notice that Time panel H has been modified compared to the one in Section 5 2 We added a Small jump step selector It determines the jump step of lt and gt buttons The other two time jump buttons lt lt and gt gt make a jump step forward and backwards of a complete time interval fromTime toTime Location controllers D are available for use only during measuring locations data visualization Color scheme controllers A allow selection of a color scheme and if it has to be discrete or continuous From the visualization panel B the type of visualization can be selected Because it has a major impact on the generated image and because it unlocks the use of some other controllers it is located at the top of the user interface panel Panel E allows the selection of additional components on the map and also has a slider that can adjust the transparency of the background map With the three check boxes the user can select if measuring locations markers visualized values and direction arrows are shown on the map City Traffic Visualization 41 CHAPTER 5 DATA VISUALIZATION
92. plit Vehicle Type Locations a Split Split MI All Vehicles M ENTRANCE EXIT M lt 56m Lim 6000 M gt 5 6m lt 12 2m M gt 12 2m My Uncategorized A lt A A 4500 Week Days O Split M vo Mm M sa _ Average M u Mr M Su C Only Avg mowe O 2 y Aggregation step Bomm w Unit y Scale Total Vehicles Automatic O Veh Hour O Fixed 5994 0 Time Calculation period 1500 From 2013 01 03 00 00 HH To 2013 01 03 ca 23 59 H R Y CA zi A a a aT Time 0 i e e ee Se a oe 00 00 01 30 03 00 04 30 06 00 07 30 09 00 10 30 12 00 13 30 15 00 16 30 18 00 19 30 21 0 22 30 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 03 01 Figure 5 25 Traffic intensity at all locations with label ENTRANCE for 03 01 2013 A split is made on vehicle class type Each polyline represent a vehicle class 10 05 2013 O All split Vehicle Type Locations Split Split M All Vehicles _ ENTRANCE s M lt 5 6m MPa My gt 5 6m lt 12 2m W gt 12 2m I Uncategorized i Week Days Split Wmo M viTn M yisa v Average Bvitu M virr Wsu Only Avg Enw OGGO 300 Aggregation step 1 hour y Unit y Scale Total Vehicles Automatic Veh Hour O Fixed 710 0 Time 130 Calculation period From 2013 05 01 00 00 H To 2013 05 34 23 59 H Jump period lt lt lt gt g
93. que number used for identification of the entry The first one is added by the database upon import of the SQL file The second one is proovided by NDW and the information there was stored in the Shapefile After step 3 we decided to keep both of them as they can be useful for future identification of the roads Column symbol does not provide any information of interest to us that is why we drop it The last column geom keeps information about the path of the road on the map After Step 2 this column is in geometry MultiLineStringZM format We decided to transform it to JSON format This has the following advantages first it will be more easy for the user to read and understand what is written in that column and second we can use libraries that are specially developed to work with JSON string which makes their reading and processing much easier After Step 3 we have the final result Figure 4 6 We name this table tmc_ line Columns gid loc_nr and geom are renamed respectively to index id and coordinates Each road is given as a series of locations through which it is passing Each location is given with its three dimensional coordinates longitude latitude hight The height of all of them is 0 so in our work we only use the latitude and longitude City Traffic Visualization 21 CHAPTER 4 DATA PREPROCESSING index id coordinates PK characte character var text 1 2001 ItypezMultiLineString coordinates 4 82176867336509 2
94. r research we did not find any entry that has any of the Uncategorized indexes that is why we decided to make the best use of them by calculating the extra vehicles using N Uncategorized N All_vehicles N vehicles lt 5 6 4 2 N vehicles lt 5 6 lt 12 2 N vehicles gt 12 2 i and save the result in the locations data table along with the information for the other indexes The resulting table GEO01_SRETIO43r is shown below Figure 4 3 periodstart periodend 15 2B 3B 4B 1A PK timestar timestamp wi real real real real real 2013 01 01 2013 01 01 15 0 g 60 45 2013 01 01 2013 01 01 1 1 1 2013 01 01 2013 01 01 36 g g 60 24 Figure 4 3 Table GEO01_SRETIO43r loaded in a PostgreSQL database and viewed with pgAdmin III These two preprocessing steps give a huge reduction in file size Putting it into numbers a 2013 year traffic intensity data for all measuring locations in the city of Eindhoven takes 53 8 GB of storage space After optimization it needs only 3 04 GB which is 5 6 of the original data and we do not have any important data loss Although NDW provides different types of traffic information the focus of this thesis is on data about traffic intensity The reason for this is that only a small number of the measuring locations provide more than traffic intensity data in practice too few to be useful NDW is constantly developing and improving so we expect that in futur
95. regated data is easier to interpret and bigger aggregation step is preferable Our application can also aggregates data depending on the step set by the user Because of that we still recommend to download raw data with as lowest aggregation step as 16 City Traffic Visualization CHAPTER 4 DATA PREPROCESSING possible because it can lead to more thorough analysis Raw data files store all available data in one big table This can make it slow and difficult to find and extract the information which is needed There is also lots of duplicated data which uses lots of memory Because of this some processing is done with the raw data before it is used by our application For each measuring period and for each measuring location we have several entries in the table rows 2 to 5 for location GEO01 SRETT043r 6 17 for GEO01_SRETI13r 18 29 for GEOO01_SRETI40 etc These entries for each location differ in the index column C field The meaning of the indexes is explained further in this section As we mentioned already each entry stores all meta data for the measuring location From here comes the first data processing step store the meta data for each location in a separate table keeping only one entry for each location The table is named PointsMetaData Table 4 2 column title column column title measurementSiteReference startLocatieForDisplayLat measurementsite Version start LocatieForDisplayLong G LocationCountryCode H i A
96. rom NDW Interface beschrijving Figure 5 4 Measuring locations linked to John F Kennedylaan The arrows on the road show the traffic direction and the dots in blue indicate the locations that measure the traffic in that direction 26 City Traffic Visualization CHAPTER 5 DATA VISUALIZATION a Nitra am NCUWIANAS h Kr DOL RUNGEKKersivelieg St rahe estel A 4 r Veldhoven VEN Ove ZWA A IN erie ae 3 Le Z go 40 WZ 155 es Nao JE am E gt JA aire DA oO gt NAO 6 ANA je CO yz Dene wt nel DN k bod p r Oti A IKnoopountl eendernende A IS he knooppunti eenderiende Figure 5 5 Labeling measuring locations in the city of Eindhoven Label ENTRANCE green is used for incoming traffic locations and label EXIT red for outgoing traffic locations City Traffic Visualization 27 CHAPTER 5 DATA VISUALIZATION 5 2 Basic controllers Following the problem description and requirements users should be able to apply filtering on the data The three main filters which were identified and applied are e Filter based on vehicle type e Filter based on measuring location or group of measuring locations label e Filter based on time Vehicle Type Locations Time All Vehicles EO01_SRETIG8 EE lt iim _ JGEO01_SRETI005r From 2013 01 03 00 00 H _ 6E001_SRETI006 gt 5 6m lt 12 2m 6E001_SRETI008 To 2013 01 03 23 59 gt 12 2m EO01_SR
97. s on y Axis of the chart is done in a different way We use a fixed number of labels evenly spaced presenting values from the 0 maa interval In our tool six labels are used Figure 5 23 We believe that this number is sufficient to give a good overview about the values used for drawing the polylines If the exact value in each point is needed then other mechanisms are available for that To determine the maz value that can be presented on the chart we use a slightly different approach than just finding our largest calculated value We apply algorithm 4 to find it and also determine the y Axis labels The algorithm calculates maz value in such a way that the interval between two neighbor ing labels has a size multiple of 5 This makes it easier to obtain a value from the chart and also produces a better looking and easier to understand image lo make it easier to identify the value and the time stamp from the lines on the graph thin gray lines were added at the background each one running from a x or y Axis label Figure 5 23 To paint the lines on the chart we need to use colors that are easily distinguishable from each other Steven Few recommends the following twelve colors red green yellow blue black white pink cyan gray orange brown and purple If more are needed then it is possible to use darker or lighter versions of the colors From these colors we already use gray white and black to draw the background and axis l
98. ses colors to show the relative speed of the vehicles on the particular stretch in the city Figure 3 4 The color used for some road is selected after comparing the average speed of the vehicles to the speed limit there Green color means that the traffic speed is closer to the limit and red mean that the speed is much lower than it Unfortunately the application does not present the velocity in numbers which can give you false idea of the traffic situation Having a central street and a highway in the same color can be interpreted as the average speed of the vehicles is the same on both places although the speed limit of the highway is 120km h and in the city 50 80km h Google Live Traffic uses data which it collects from devices running Android OS smart l https maps google com 8 City Traffic Visualization CHAPTER 3 RELATED WORK Figure 3 3 a Aggregation techniques to visualize the vessel density in the area of Rotterdam harbor The author uses colors and illumination to show the variation of the density levels 23 b Detail lenses to present more details about the route Lenses can be used to display more and varied information like zoomed images or traffic numbers 19 essa RJ Wandelpark Eckart A pad AP ee Muschberg Geestenberg i xa per Do akkers West AO Lakerlopen Berckelbosch be gt e Irisbuurt ES a gt lt A Burghplan Hotel i A rs Gijzenrooi Gestel
99. similar way First select a road Second select a direction for which you want to bind markers by pressing either Positive or Negative radio buttons Select the markers for the chosen direction Press the Update button If you want to remove all markers assigned to some direction of a road select the road and the direction and press the Clear button This removes all markers from the corresponding table Figure B 12 Deleting unnecessary objects Top image selected objects are painted in red for the roads and blue for markers Bottom left image the selected objects are deleted Bottom right image only selected objects are kept City Traffic Visualization 13 APPENDIX B USER MANUAL If you want to see the locations bound to a road click on the road with the left mouse button The tables and the fields on the right are automatically filled with the road data The Delete option allows you to remove objects from the map First select objects by clicking on them or drawing a selection rectangle around them Selected roads are painted in red and selected markers in blue Pressing the Delete button removes the selected objects from the model Pressing the Keep button deletes every object but the selected one Figure B 13 B 1 4 Views Menu s SR City Traffic Visualization File Data Map Editor Views El Figure B 13 Views menu Buttons from left to right map view charts view 1 Map Vie
100. sualization techniques that visualize the vast amount of city traffic data which can be used to see how different types of policies would affect the traffic flows of both people and freight This thesis is structured in the following way In Chapter 2 we analyze the problem and set the main goals for this project Chapter 3 presents some solutions that have been developed in the area of visualizing data on a geographical image and a discussion of some existing applications dealing with traffic data Chapter 4 describes the shape and format of NDW data and how we process it to suit our needs Chapter 5 introduces our visualization solutions and gives details about choices made Finally we combine everything in one system Chapter 6 City Traffic Visualization 1 Chapter 2 Visualization Requirements The current project is intended to support the researchers from the Smart Logistics Re search Lab in Operations Planning Accounting and Control OPAC group at Eindhoven University of Technology As stated in the name the OPAC group is focusing on the plan ning and control of operational processes both within the manufacturing and service sectors These include 1 Manufacturing Sector Production control distribution transportation warehousing and retail processes 2 Service Sector Maintenance management health care and public transport Smart Logistics focuses on the efficiency and effective alignment of planning and sch
101. surementSiteNamel AT publication Time W measurementSiteName2 AU deducedNoTrafficNnntes X measurementSiteNumberOfLanes AV carriageway Table 4 1 List of the attributes that can be found in the traffic data files The columns that represent metadata for each measuring location are shown in blue oo Po MM A Figure 4 1 shows a short example of the data By exploring the rows we can see that they are first sorted by measuring period columns D and E and after that by measuring location name column A The periodStart D and periodEnd E columns specify the aggregation period of the measurements NDW uses this aggregation step to create the measure values columns Q R and S Hence by looking at the data we can answer for example the question How many vehicles have passed during the period aggregation step but not At what time has each vehicle passed This creates some issues with the accuracy and consistency of the data itself E g if we have 10 vehicles which have passed through the measuring location in a period of 5 minutes this cannot be interpreted as if 2 vehicles have passed every minute The developed application was made to work with this limitation but it should also be kept in mind by the user itself To solve this issue we can use a smaller aggregation step the smallest one allowed by NDW is 1 minute but this increases linearly the size of the downloaded data Ofcourse in some cases agg
102. t gt Time 00 00 02 00 04 00 D6 00 08 00 10 00 12 00 14 00 16 00 18 00 20 00 22 00 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 02 05 Figure 5 26 Traffic intensity at all locations with label EXIT for May 2013 A split is made on days of the week Each line represents traffic data for one day Thicker darker lines show the average data for the same week days Positioning the mouse over a line results in highlighting it showing a box with information about the line Positioning the mouse over a point with a value results in showing the value there City Traffic Visualization 49 CHAPTER 5 DATA VISUALIZATION Number of vehicles Chart type Lines Split Location From 2013 11 1500 00 To 2013 11 1523 59 AggStep 30min Vehicles lt 5 6m Locations GEO01_5RETI043 GEOO SRETIO13 GEOO1_SRETIOSS1 GEOO1_SRETI681 GEOO1_SRETI441 GEOO1_SRETI311 1750 1400 1050 700 350 Number of vehicles Chart type Lines Split Location From 2013 11 1500 00 To 2013 11 1523 59 Agg Step 30 min Vehicles lt 5 6m 10000 8000 6000 4000 2000 09 00 10 30 1 15 11 15 11 1 00 13 30 15 00 11 15 11 15 11 Lato Locations ENTRANCE EXIT Ring Clk Ring CoClk Time a Each line presents a data from one measuring location
103. the type of visualization 5 3 Visualization of traffic on a map Visualizing traffic data on a map is one of the main requirements which we have for this project We use the already created city model to create views Before we create any visu alization we need an appropriate map We decided to use the standard layer Open Street Map This layer provides information about roads buildings area borders etc Every dif ferent type of objects has it own color which makes the map fast and easy to understand and follow Being able to distinguish the separate objects like specialized buildings for example can prove useful in the later analysis of the data All map based views are designed to show the spatial properties of the data For that we use the location of each measuring station Before generating any image we have to make the appropriate calculations based on the user preferences and applied filters We use algorithm 1 to obtain the final results and then we keep the numbers for each location Algorithm 1 Pseudo code for generating values for each location 1 Input List of locations L List of vehicle classes V Time period T tpegin tend 2 and flag vehiclesPer Hour 3 Output Value result 4 for location in Z do 5 Obtain data D for location l vehicle classes V and time period T 6 sum 0 e for line d in D do 8 for Vehicle class v in V do 9 sum sum d Note d is the number of vehicles for class v in lin
104. the week combined All these four cases have their own specifications which will be mentioned below Algorithm 2 Pseudo code for generating values lists making split on vehicle class 1 Input List of locations L List of vehicle classes V Time period T tpegin tend 2 aggregation step tAgg and flag vehiclesPer Hour 3 Output List of list of values vLists that are to be visualized 4 List vLists 5 for Vehicle class v in V do 6 List buf List T for location l in L do 8 Obtain data D for location l vehicle class v and time period T 9 time tbegin sum 0 List resList 10 for line d in D do 11 if dtime tbegin tbegin tAgg then 12 sum sum d Note d is the record for vehicle class v in line d 13 else 14 if vehiclesPerHour then 15 result sum total Lines 16 else VE result sum totalLines x tAgg 60 18 Add result to resList thegin tbegin tAgg sum 0 19 Sum elements from resList to the corresponding elements from bufList 20 Add vehicle class information to list buf List 21 Add bufList to vLists Calculating the values lists for splitting on vehicle class is done using algorithm 2 It calculates one list for each individual vehicle class selected In this case we sum the data from the individual locations together The extra information that is added to the list line 20 is later used for our visualization purposes e g determine the color of the line tAgg is the aggregati
105. tion of the marker To combine two paths together select the Combine option Then select the two roads that you want to combine Selected roads are drawn in red Click Combine button Figure B 8 2 Cut ii Combine Combine Figure B 8 Edit Roads function In the top image road 1 has been cut into two parts road la and 1b In the bottom image road 1b has been combined with road 2 The Create A Label functions allows binding of several markers into a group by assigning them the same label To create a new label select the NEW LABEL option from the combo box Then write the name of the new label in the Name field Click on the small square next to Color to select and assign a color for this label Select locations that will be tagged with the new label Selecting locations can be done by either clicking on them or holding the left mouse button and draw a selection rectangle around them The selected locations are painted in blue Press Put Label to assign the label to the locations Figure B 9 If you want to assign more locations with an existing label first select the locations After that select the label from the combo box and press Put Label To see the locations with a given label simply select the label from the combo box and press Show Markers Locations with that label are painted with the color assigned to the label 70 City Traffic Visualization APPENDIX B USER MANUAL To remove a l
106. tton allows creation of a new label and allocating it to markers 5 6 See Label s Markers button shows the markers assigned with the selected label 7 Bind Markers To A Road binds a selection of measuring locations to a selected road 8 Delete button removes markers and roads from the model The Move Marker option can be used to adjust the display location of a measuring location marker Figure B 5 Changing it is done by clicking and holding on the marker with the left mouse button and while holding it move the mouse around to drag the marker to the desired position The selected marker is painted in blue color The gray ghost point on the map shows the original marker position The marker and its position are connected with a thin gray line If the user wants to return the marker to its original position he can press the RESET button Reset location RESET Figure B 5 Map marker with adjusted display location The blue point represents the new display location the connected gray point shows its original location Edit Marker Data allows the user to alter the original meta data for each marker Figure B 6 First select the marker by clicking with the left mouse button on it The marker gets colored in blue and its data is loaded to the text fields on the right Make the required adjustments of the data and then save them by pressing the SAVE button WARNING By pressing SAVE the data from the fields is saved directly t
107. ualize data based on label or individual locations Radio buttons 2a 2b 2c and 2d allow the selection of the attribute on which data should be split 2a splits the data based on vehicle type and location 2b on vehicle type only 2c on location only and 2d on day of the week Button group 4 has the same function as in the map visualizations interface In charts the colors assigned to each class can determine the color of the drawn lines or bars To change the color of the class press the red square and select a color from the color swatch Button group 5 gives the option to select for which days of the week to see data fig ure B 18 By default it is possible to select several days The four buttons in the Week Days panel have the same function as the ones in Vehicle type panel select one select all deselect all and invert selection City Traffic Visualization 19 APPENDIX B USER MANUAL The controllers in panel y Scale 6 allow you to set the maximum value which is visualized on the chart Based on the type of chart and grouping selection the Locations panel contains different controllers versions A B and O Version A is available when we have lines chart and labels Table 7 shows all predefined labels and from there the user can select for which labels he wants to visualize data Version B is available when we have selected Points from combo box 3 This panel and controllers are the same as the one in map visualizations user inter
108. visualizing traffic data over a geographical image and visualization using line and bar charts The first type of images allow the exploration of the spatial property The user can see the traffic situation in a whole city for any period in time Charts assist the user in understanding the vast amount of data Multiple controls are provided to change the dimensions such as temporal properties vehicle class location etc Finally we combine everything in a multi user system City Traffic Visualization 111 Acknowledgements I would like to thank my supervisor Prof dr ir J J van Wijk for his guidance during this project His comments feedback and countless advices on the prototype and the report helped me to improve my ideas and achieve my goals I would also like to thank Mark Stobbe for his help with the technical part of this project His help with Java development and databases proved to be more than helpful I would also like to thank all my friends and colleagues for their feedback on my work and tor their help and advices on the design of the images in this paper and the visualizations Finally I would like to express my gratitude to my family for providing me with the opportu nity to come to Technical University of Eindhoven and for their constant support during the time I spent here City Traffic Visualization V Contents Abstract 111 Acknowledgements Vv Contents vii 1 Introduction 1 2 Visualization Requirements 3 al ODEN e
109. w button allows visualization of geographic traffic data on the map 2 Charts View button allows visualization of temporal traffic data with charts B 1 5 About Menu Welcome About City Traffic Visualization Alt A Figure B 14 About menu 1 Welcome opens the welcome screen 2 About City Traffic Visualization Opens a window with information about City Traffic Visualization 74 City Traffic Visualization APPENDIX B USER MANUAL B 2 Creating a model A city model is the base on which our visualizations are build It is required for the proper functioning of City Traffic Visualization The creation of a model is done with the help of the Map Editor menu options and is recommended to be done in the following order 1 2 Load a connection string file if not done so From File menu select Create New Model If the user wants to continue working on an existing model then select Open Model and load the model file With the help of Delete option remove all roads and markers that are not going to be part of the model This reduces the clutter of the generated image If needed use Draw Roads and Edit Roads options to shape the road network in the model More accurate road network makes more accurate road traffic visualization Select Move Marker option to adjust the display position of the markers The display position is important for the accurate visualizations on the map Try to position the markers close to th

Master`s Thesis

Contents

Download Pdf Manuals

Related Search

Related Contents