Analytics User Manual
Contents
1. eene 26 CREATE A CONTINUOUS VARIABLE 25 55 2 Freguent y rire UENIRE ME INN ONT UE 2 GENERALIDESCIIPTION s NU PITE ES 2 SINGLE VARIABLE CATEGORY FREQUENCIES eese eene nennen nennen nnne nnn rennen tears 27 TWO WAY CROSS FREQUENCIES 27 Mene I SIS E TOTO 27 GENERAL DESCRIPTION vccnscsieaststagenesnceasadepetcatenpensbelenihaweapeciiadershaiteaieiabateroionntajequadepotmsasderaautageianeetede 27 ANA Tm 28 MB SUO E A 28 ipe d Ra A HI ECT 29 eis a MB SIR e e 29 SELECT A BINARY VARIABLE TO PROFILE eese nnne nnne nennen nnns senis 29 CHANGE P VALUE M 29 Rapid I nsight Analytics User Manual FRO LE essere 29 VP FH 29 PRORILE REPOR 29 APORTO EA CEE 30 PO 302. 30 FERED IOI Mid M 30 MISSING VALUES DEFADEFDTHANDEING s edoaaiso des De Rad ER ED TREAT d eU RAD nS 30 P VALUES FOR 31 PIV WN 31 31 BEGIN REGRESSION ANALYSIS aone EEan a UFU EEEE EET E 33 Building a Model to Predict 33 GENERAL DESCRIPTION Mr aiiai 33 Be AES rr 33 T eerie tits ses ry nce VE EE 34 MEMORIZE MODELS C 34 REGRESSION RESULT c Hm 34 MIS PIS MIO iios eT 34 COMM SIV OG CIS occ pcs asia cUm VM teme EIU bes RN I INN P n dq UE EPI TUE 35 GENERAL IDES CRIP TION cerris DU DUE QUEM FA EE KOC dT Se uU CURAR 35 TERN EE 35 GENERAL DESCRIPTION seis sesssas aevo bava ax E ee pnr ERESFR ER RERERa Fr ERES FERTESETRAY e ETE NVRREERERE Yu oT ENERE E REY SERRA 35 SELECT A MO
3. Note window Rapid Insight Analytics User Manual SCORING MODULE Select and Load a Dataset GENERAL DESCRIPTION The Scoring module can provide a predicted Y variable for each X variable given any model It also evaluates the model based on how closely each predicted Y variable correlates to the actual values of the Y variables SCORE TO We have the option of scoring to a flatfile or database column by selecting the appropriate button option in the Score to box SELECT DATASET Click Select Dataset and choose the type of file you d like to open from the tabs on the left side of the screen Find the correct file path and click OK to open SELECT SCORING MODEL When your dataset has loaded click Select Scoring Model and choose a rism Rapid Insight Scoring Model file to evaluate Just Score the Dataset Score the Dataset and Validate the Model Choosing the Just score the dataset option presents only a column with a predicted Y value appended to your original table or dataset as output Choosing Score the dataset and validate the model will generate a chart below the drop down menu displaying the decile analysis or cumulative list decided by the toggle buttons to the right of the drop down menu of the model for predicting the given Y variable Output to Flatfile n d Here we have the options of Output Decile Output Percentile or Don t Output Decile Info in the drop down menu Choosing to out
4. TICS modeling Analy predictiv User Manual d e _4Rapid Insight simplifying data analytics Copyright 2012 Rapid Insight Inc All Rights Reserved Rapid I nsight Analytics User Manual TABLE OF CONTENTS ieu ique M MA LS DoOCUInefitus eni met uten OOT IMPIDE LIMITI EHE LMM NNI UI 5 Hardware Requirements ccccscccesecceseccceecceeceeencceeueceeeceeeeceeeeseeeuceseeeseueceegeceeeeseeeneees 5 Additonal RE OUNCE emerat dmt EUM ME EER 5 PROGRAM ELEMENTS sisi 0 Analytics MOU t eee eee ee eee DEINEN EDI UM 6 SC OFM MO dU T EM 6 VRE Variable Reduction Engine Module eese 6 ANALYTICS MODULE NERVES FETA VE qVAT RARI up OR FOTO dkua i Seleer and Load a Dataset uncos uccpnied faxis E Keds 7 GE NER AIDE SCR PO Bl T 7 LOAD SAVED ANALYSIS AND DATASET ccccccsecsccscceccecsecsececcecsessecsececcessesseseccecessesseseceessesseseeeesoesass 7 OPEN DATASET 7 SELECT DATASET TO 7 LOAD VARIABLE sorreraren 9 LOAD SAVED ANALYSIS ON CURRENT DATASET cesses nnnm mener
5. fill in the New Variable Name field and click the Create this Variable button You can any of the new variables created by selecting a name from the New Variables Created window The formula for this variable will then appear in the Current Variable Definition window USE MEMORIZED MODEL S EQUATION TO CREATE A VARIABLE After selecting the Use Memorized Model s Equation to create a variable button memorized models will appear in the Select Variable to Transform window instead of variables Select one then if desired change elements of the model formula within the Model Equation window before providing a name and saving the new variable Rapid Insight Analytics User Manual CREATE A CONTINUOUS VARIABLE After selecting the Create a continuous variable from text button select the text variable to transform rename the variable in the New Variable Name and click Create this Variable Frequency Analysis GENERAL DESCRIPTION The Frequency Analysis tab looks at the frequency of occurrence for binary and categorical variables SINGLE VARIABLE CATEGORY FREQUENCIES By clicking the Single Variable Category Frequencies button the counts and percentages for each variable selected are displayed Click Submit to view the Boe eee Mei Maier selected variables values counts and percentages in the Gender window below the variable lists Users may also choose to Female 1007 40 0 view th
6. Logistic regression is used when the Y variable is a binary The resulting model looks to predict the likelihood of an event to take place OLS Regression Ordinary Least Squares OLS Regression is used when the Y variable is continuous Users may select the Logistic regression method to model continuous variables though this is not recommended The resulting model looks to predict the rate or quantity of an event MISSING VALUES DEFAULT HANDLING Missing Values Default Handling The Missing Handling column identifies any missing values within each variable If missing Exclude Entire Record values are found an option must be selected from the drop down menu Default which is defined on the Control Panel tab Fill with Rapid Insight Analytics User Manual Variable Mean or Exclude Entire Record It is important to keep in mind that excluding multiple records can greatly reduce sample size for analysis do so with caution Fill with Variable Means Clicking the Fill with Variable Means button will fill in the missing values for each variable with the average value for that variable Exclude Entire Record Clicking the Exclude Entire Record will not replace a missing value with any other value rather it will throw out that record for all variables so that it will not affect the analysis P VALUES FOR STEPWISE P Values for Stepwise The P Values for Stepwis
7. Math Correlation Matrix displays the correlation 5 chnlarshin coefficients for each variable pair The correlations range from 1 to 1 and indicate the degree of association between two variables A positive value implies a positive association while a negative value implies an inverse association The closer each value is to 1 or 1 the stronger the degree of association Select All Clear All Select Variables Select All to see all of the variables and their statistics at once Clear All to clear all variables from the table or individually Select Variables by clicking the checkboxes next to each one Export to Excel The Correlation Matrix can be exported as a spreadsheet to Excel by clicking the Export to Excel button below the list of variables Rapid Insight Analytics User Manual Profiling Analysis GENERAL DESCRIPTION The Profiling Analysis tab compares two groups of a binary variable to determine how statistically different they are SELECT A BINARY VARIABLE TO PROFILE Select a Binary Variable to The select a Binary Variable to Profile window lists all of the Profile binary variables in a dataset After highlighting a variable create names for the binary variable s two values in the Profile Work Area CHANGE P VALUE a ae ee Change p value The Change P Value drop down menu gives the ability to change the p value according to how statistically significant the relationship betwe
8. variable as Y axis The 96 of Sample button shows the percentage of the sample square root each category represents as Y axis The Both button shows the means or square proportions as shown with the Means Proportions o aE button as well as a blue line representing the percent of Of Sample sample that fell into each category These options are available for a categorical Bath or binary variable X variable Means Proportions of Sample Both cubed root cube View Chart Data Click the View Chart Data button to see the sum mean min max and count for View Chart Data each point or bar used in the chart We are able to save this as an Excel file by clicking Save A aet ies M Anaiyees Corm Stadents ce Anais A Die Ortons epot Mindow iiy Analyzing Athlete Flag AMALYSIS VARIABLES RELATED VARAABLES Automated Mining GENERAL DESCRIPTION The Automated Mining tab lets us take advantage of Rapid Insight Analytics ability to quickly evaluate all variables against the Y variable to find any statistical relationships that exist In order to accomplish this a Y variable must first be chosen in the Control Panel tab ANALYSIS VARIABLES The Analysis Variables column lists all of the variables from the Control Panel tab The Y variable is at the top of the Automated Mining tab and will not be included in this list WON ELATED p0 Rapid Insight Analytics User Man
9. 650 1 1 n Student ID TEES Male 460 ME 0 Legacy 2 400 1 0 11 tiasa Male 700 NH D Legacy 2 300 1 n 1 ete Female 420 NH D Non legacy 2 350 1 1 0 18S GELS Female 460 NH D Non legacy 2 300 1 1 0 I Male 520 M 1 Legacy 2 500 1 0 0 AUTOMATED LONE SESS Male 520 VT D Legacy 2 450 1 1 0 MINING 16 eiza Female 450 M 0 Nor legacy 2 900 1 1 0 iG Male 600 RI D Legacy 2 300 1 n n LB EAE Male 620 NH Legacy 2 550 n 1 n IEEE Male 600 NJ Legacy 2 850 1 1 1 Female 600 CT 0 Non legacy 2 450 1 1 0 21 225 Male 500 0 Legacy 1 600 1 n n VEEE VP Male 560 M D Legacy 2 000 1 n n Sort Ascending Sort Descending Reset Sort Order Sort any column by highlighting it and clicking on the green up v ra arrow above the grid to sort ascending or the green down arrow above to sort descending Reset the sort order by clicking the green left pointing arrow To sort by multiple columns drag the columns so that they are next to each other in the order you wish to sort Click the first column hold the shift key down and click the second column then click either the ascending or descending sort button Decile and Take Means By Selected Column To find deciles and take means of each decile of any select column highlight the column and click the Decile Mean button above the grid To return to the main grid press the green reset sort order arrow Decile and Sum By Selected
10. Name field and click the Memorize Model button This retains the model for recall later It also makes it available in other areas in the program such as the Comparing Models and Create Variables tabs REGRESSION RESULTS The Regression Results window is located below the variable lists It displays the model s variables their coefficients standard errors t values and p values All of the regression analysis steps taken to obtain the current model are also available by expanding the Model Steps section The Diagnostics section when expanded provides objective measures evaluating the quality of the model MODEL DECILE CHART A decile column chart is automatically generated when a model is produced It s a graphical representation of the results of the model having been tested by being applied to the Holdout Sample records The red bars depict the test results A set blue bars depicting the predicted outcome will appear when the Show Actual and Predicted option box that appears above the graph is checked There are three separate decile views Decile Analysis Decile Lift and Cumulative Lift Decile Analysis The decile analysis divides the total population into ten equal bins deciles The bars on this chart represent the average Y value for each bin Decile Lift The decile lift Shows the percentage increase in accurate prediction by using the given model Cumulative Lift The cumulative lift shows the cumulati
11. at a minimum contain the same variables used to originally create the analysis The program will automatically open to the Control Panel tab Rapid Insight Analytics User Manual CREATE ANALYSIS After clicking the Load Variables button the Create Analysis button becomes available to begin a new analysis Once clicked the program will automatically open to the Control Panel tab ANALYTICS MAIN MENU GENERAL DESCRIPTION The main menu of Analytics allows us to load save and export analyses control the physical appearance of Analytics change default data settings launch the Report bar toggle between various windows and access the Help menu FILE MENU File Options Reports Window Help Load Analysis Load Analysis Ctrl O After loading the selected dataset the Load Save Dataset Ctrl S Analysis button is available under the File menu Export Dataset Ctr E If you have previously conducted an Analytics f analysis on the uploaded dataset you can apply the Close Dataset analysis by selecting the appropriate ria file here Close A Save Analysis New Analysis Ctrl N To save all parts of the current analysis click Save Analysis under the File menu at any time Exit Export Analysis Because we have the ability to set filters cap outliers and add calculated variables to a dataset in Analytics sometimes it is necessary to export this data to a csv file To do thi
12. button to view them To view the effect that outliers have on each chart choose View Effect on Chart to view a new chart which does not include the outlier values Note to actually exclude outliers from the analysis set a column filter on the Control Panel tab Find Outliers Print Chart Print out individual charts by clicking Print Chart Print Chart Show Missing Values E EE Check Show Missing Values to see a pie slice or bar for the missing values Show Missing Values OTHER Slice Threshold Slider Categorical charts with a number of very small slices may be modified to combine the small slices into one large slice by sliding the Other Slice Threshold slider bar The eel esl slice that contains the combined smaller slices is always yellow and does not show a count or percentage To see a pie slice or bar for the missing values check the Show Missing Values box Multivariate Analysis GENERAL DESCRIPTION The Multivariate Analysis tab allows us to analyze two or more variables simultaneously illustrating how pairs of variables may be correlated Rapid Insight Analytics User Manual A Ra pid Insight Analytics Current Students csv alysi A File Options Reports Window Help CONTROL i PANEL Bl X ans Proportions VARIABLE Y axis Variable x ais Variable View By Exclude Categories with Tr STATISTICS Campus Housing Flag vB TEENS x values Li
13. nnns 9 CREATE ANALYSIS E 10 ANALYTICS MAIN susan navon ons arinari isnat 10 GE INE RAD es CRIP TON ee S 10 EN esters m 10 ene saumiusBm 11 REPORTS MENU e 11 CONOP A BL setate E CM ENRN C VEM MEME E UAM DEMENS EOS DUE 12 GENERALDESCRIPTION m 12 VARIABLE CONTROL PANEL 12 Y VARIABLE 13 EA E E M 13 8 51 5 155 14 GENERALDESCRIPTION 14 VARIABLE STATISTICS GRID ern EAI 14 VARIABLE STATISTICS LEFT SIDEBAR cccceccscceccecsecsecsececeucsecseceecececseseececeessesseeeeeecee
14. 1 110 00 570 00 460 00 continuous STATISTICS Campus Housing Flag Gender n a n a n a n a n a categorical Campus Visit Flag H5_GPA 2379 03604 1100 3950 2 750 continuous Days between app date and t Legacy Indicator n a n a n a n a n a categorical Gender SAT Math 588 10 8266 40000 800 00 400 00 continuous 1 000 binary n a categorical 2519 999 413 continuous i MIHS_GPA Scholarship Flag 09270 n a 0 1 000 Legacy Indicator State n a n a n a n a WI SAT Math Student ID 48559944 29111451 352 00 999 765 e omoocce UNIVARIATE Scholarship Flag ANALYSIS State Student ID MULTIVARIATE ANALYSIS VARIABLE STATISTICS GRID This grid contains information about each variable s mean standard deviation minimum value maximum value data type coefficient of variation number of missing values number of observations number of distinct values range and whether a variable is text or not Mean The mean is the average value of all of the observations of one variable Std Dev standard deviation The standard deviation represents the spread of data around its mean A higher value indicates a more spread out data set than a lower value Min minimum The minimum value represents the smallest observation found Max maximum The maximum value represents the largest observation found Coeff Var coefficient variation This represents the ratio of the standard deviation to the mean The larger this number is the la
15. ABLE RELATIONSHIPS MAIN MENU The Variable Relationships Menu allows us to view charts of each variable involved in Analysis We can toggle between those variables deemed related or unrelated by the Automated Mining tab to see each variable and the relationship it has with our designated Y variable Variables Menu View Related Unrelated Variables View Chart Report Export The View Related Unrelated Variables lines allow us to View Related Variables toggle between graphs of variables that the Automated Mining tab has decided are unrelated to our selected Y View Non Related Variables variable and those that are related to our Y variable We can gain more insight into these relationships by also selecting a View By variable from the drop down menu L View Menu Statistical Test Chart Report Export The Statistical Test line allows you to see the statistical test values TE 1 for the variable shown Depending on the type of variable selected Chart Data the output of this feature does change Generally a Z or F value is Lift shown along with any pertinent information Statistical Test Values on Chart v 3D Chart Category Data k Rapid Insight Analytics User Manual Report Menu Create Report on these Variables The Create Report On These Launch Report Bar Drag n Drop Graphs Cha
16. ANALYZE Relationship between Campus Housing Flag and Athlete Flag Proportion MULTIVARIATE ANALYSIS AUTOMATED MINING CREATE VARIABLES ANALYSIS FREQUENCY MEANS ANALYSIS E wr ge ANALYS E i REPORTS PROFILING ANALYSIS Campus Housing Flag Rate 5 5 e e 5 o MODEL NOTES MULTIVARIATE UPPER TOOLBAR CONTROLS This toolbar modifies the charts that the Multivariate Analysis tab creates Control whether charts are created from the sums or means of variables It add labels to charts to make them easier to read changes their formatting to make them more aesthetically pleasing and print charts Use Lift Index on Y Axis To view the lift index for each entry on the Y axis rather than the Y value click the Use Lift Index button The lift index shows the Y axis indexed to one and for any value that falls above or below 1 0 we know that the variable mean for that combination of data is above or below the population mean Show Counts By clicking the Show Counts button the user can Relationship between Campus Visit Flag and Athlete Flag view counts as a label for each bar or point on our chart Athlete Flog 3 D Toggle Button Clicking the 3 D Toggle Button allows users to B switch between 2 D and a rotating 3 D view of a chart This button is available only when the X axis and View By variables are both either binary or categorical To
17. Column To find deciles and take the sum of each decile of any select column highlight the column and click the Decile Sum button above the grid To return to the main grid press the green reset sort order arrow Search for any data value by clicking the Find button or using CTRL F The find window gives the option of searching by just the column and row of a highlighted cell by clicking the Column Row Only button or searching the whole grid by selecting the Everywhere button Searches are done by row or by column depending on which option is selected ANALYZE DATA LEFT SIDEBAR The left sidebar displays only the selected variables The table adds or removes variables from view making it easy to place variables next to each other to see how they compare Select All Clear All Select Variables c Rapid Insight Analytics User Manual Use Select All to see all of the variables at once Clear All to clear all variables from the table or Select Variables individually by clicking the checkbox next to each one Export to Excel The results of a decile aggregation of the records can be exported as a spreadsheet to Excel by clicking the Export to Excel button below the list of variables Export ta Excel Univariate Analysis GENERAL DESCRIPTION The Univariate Analysis tab graphically displays only one variable at a time View the distribution of any variable in the dataset by simply
18. DEL MODEL 5 35 t A 36 ea gU n NU RESCUE UR ROM M ras 36 DOCUMENT EDITING CONTROL CPP TRETEN 36 NOLO c 37 GENEHRALIDESCITIPTION ceirnin A EE VIRI UM ROS IIR lesa aoaaindan ETET 37 SCORING MODULE wecessinsacccnsecsssescacsncedsnnsescsntencsnnsscvetesererissaipecetseaternsne DO Select and 556 858 1856 38 38 m 38 A M M 38 SELECT CORIN G MODE T 38 CONTACT a G Rapid I nsight Analytics User Manual ABOUT ANALYTICS Analytics is a powerful desktop application for analyzing analytic data sets and building predictive models Predictive modeling is an analytic approach that identifies patterns and relationships between data variables These relationships are statistically validated and then used to build analyti
19. Note that selecting Windows Desktop directs the palette to match whatever the Windows Desktop s colors are set to Dataset Evaluation Count The Dataset Evaluation Count drop down menu allows us to change the number of records taken into account to determine the datatype of each column variable 5 000 is the default number of records used to determine datatype note that the larger this setting is the more precise the datatype evaluation will be though increasing this setting may take more time to load Manually changing a column s datatype can be done by using the Define Format option when selecting the dataset or selecting a datatype from the Type drop down menu on the Variable Control Panel Dataset Query Timeout The amount of time selected in the Dataset Query Timeout drop down menu corresponds to the amount of time Analytics will allow a server to take before responding If the server does not respond within the specified amount of time Analytics will perform a timeout Load New Dataset Columns Into Analysis The Load New Dataset Columns Into Analysis option grants Analytics the ability to load in any columns that have been added to the original data source post analysis into the current analysis REPORTS MENU Window Help Launch Report Bar Drag n Drop Graphs Charts Launch Report Bar Variable Statistics The report bar gives us a handy place to drag and drop any charts or tables we wish to capture anytime d
20. Rapid Insight Analytics User Manual stop the auto rotation click anywhere on the chart and the chart will freeze To see other views drag the scroll bars on below and to the right of the chart Angle X Axis Y Axis Annotations A These options angle the X axis and Y axis labels to improve legibility Sort Bars by Height The Sort Bars by Height button is available if the X variable is binary or categorical This button orders the bars from largest to smallest left to right Print Current Chart Print the chart in view by clicking the Print Current Chart button Chart Y Axis with Means Sums Choosing either the Chart Y Axis with Sums or Chart Y axis with Means button determines how the Y axis will be distributed MULTIVARIATE LOWER TOOLBAR CONTROLS The lower controls on the Multivariate Analysis tab select and sort which variables to compare Users can also exclude categories and view these charts by their means or the percentage of the sample that they represent Data records represented in each chart may also be viewed Y axis Variable To look for relationships between variables manually first select a Y axis variable to compare with multiple X axis variables This way users can visually compare all variables against each other The Y axis variable can only be continuous or binary If it is binary the Y value is expressed as a rate X axis Variable Choose each X axis variable to compare
21. T I 21 RELATED VARIABLES III 22 Sl dci peu UAB 22 CHANCE NS TE 22 PA OVINE AI ever c A 22 VIEW RELA NON T EEE AA 22 Variable Relgtobs U 22 GENE TO 22 VARIABLE 5 5 22 VARIABLE RELATIONSHIPS MAIN MENU eene n nnn nennen nnn nnne nnne sre rne 24 wenn E 25 GENERA CODE CIP TO IN orerar T E EEE 25 SINGLE VARIABLE OPERA TION S 25 CREATE A BINARY VARIABLE ccsccsccsscceccsccsecseceeeeeeeceeessesseceeeeeeseeeeeseesesaeeaecseceecseeseeeaesaeseecaeeeees 25 sce T 26 MULTIPLE VARIABLE OPERA TIONS ussusseabaq tusure ER FUsEe QUT dE IESU KE NET D EECIR pas Esc ERE UY VERUS rtp Kern AS FO TENER CEN UNE 26 USE MEMORIZED MODEL S EQUATION TO CREATE A VARIABLE
22. alytics User Manual Data Type Categories The broad data categories Analytics A Select Dataset toOpen acc E hand edge of the window Clicking a Bap Insight Crime Data csv i indo Other Electrode History YTDO05 Q4 only csv button will allow the user to search for files or tables of that corresponding data type These data categories are Text File e RAPID INSIGHT DATASET This accesses a proprietary file type rids provided by Rapid Insight e TEXT FILE This type covers asc CSV tab and txt data files In addition to selecting the file the user may specify the way in which the file is read by the program Options include whether the fields are fixed width or delimited the type of delimiter if any and whether the first row contains headers The user may further specify the format of the file by selecting the Define Format radio button and clicking OK re Delimited Fixed width Define Format Saved Jata Link m in fi Data Link Headers in first row Cancel e DATABASE TABLE This section is divided into 4 data type sub categories o FILE BASED This allows access to file based database types such as MS Access FoxPro and Paradox as well as formatted data types such as XML o OLEDB This section provides access to server based datasets such as MS SQL Server and Oracle using an OLE DB data driver The user must specify details of the connection such as the datab
23. ase name and user credentials o ODBC This section provides access to server based datasets such as MS SQL Server and Oracle using an ODBC data driver The user must specify details of the connection such as the database name and user credentials o View Query This option is currently unavailable e EXCELFILE Files using either the older xls or newer xlsx formats are both accessible e SAVED DATA LINK This accesses data from files using the extension udl Data Location This field allows the user to specify the location of the data source on the local machine or on a network drive Data Type Options Data type options appear in the upper right of the window under a drop down control It allows the user to define the specific file extension 5 Rapid I nsight Analytics User Manual Matching Data Files Tables This field displays the files tables that meet the data type and data type option in the selected location The user may select the dataset to open from this list by clicking on it then clicking the OK button LOAD VARIABLES Once a new dataset is selected using Open Dataset the following fields and controls are enabled Sampling This control is only active prior to clicking the Load Variables button It allows the user the option of analyzing just a subset 1 to 100 of the records originally found in the dataset If the Random Order Rows box is checked the Load entire dataset records t
24. ased Note What If functionality makes assumptions regarding causalities in the data and for this reason results could turn out to be much different than projected Rapid Insight makes no guarantees regarding accuracy of these projections either implied or explicit and this tool is to be used at your own risk SELECT A MODEL MODEL VARIABLES Explore what if scenarios using any saved model by selecting it from the Select a Model window The Model Equation window displays the established formula for the selected model To perform a what if analysis select any of the variables in the Model Variables window and use the slider control to change its value The start value for any variable is its mean The graph of the variable to the right will change to reflect the variable s new mean as one or more a slider controls is repositioned Rapid Insight Analytics User Manual Reports GENERAL DESCRIPTION The Reports tab allows us to bring all of the information we ve gathered into one report to be shared Here we can import previously saved reports or generate a report by using content saved to the Report Bar DOCUMENT EDITING CONTROLS Here we can import previously saved reports or generate a report by using content saved to the Report Bar New Clicking the New button generates a new blank report Open To open a Saved report click Open and select the report Save Clicking Save will sav
25. be in the form of a single structured data file such as an Excel or CSV file or a single table in a database such as Oracle or MS SQL Server All data to be used must be contained within this dataset no imbedded or implied relational links will be recognized Which data types are available is often dependent upon whether the 32 bit drivers or providers for that type are installed on the computer Drivers OLE DB ODBC ADO net or native drivers are freely available for download from the corresponding manufacturer Data types Analytics regularly supports include e Access e Paradox e DB2 e Rapid Insight Dataset e DSN ODBC Connections e SAS e Excel e SPSS e FoxPro e SQLServer e MySQL e Text CSV e Oracle e UDL e OtherDBMS e xBase e OtherFile e XML Load Saved Analysis and LOAD SAVED ANALYSIS AND DATASET Jataset Dese Users selecting this option will identify a saved Rapid Insight analysis file ria and open it along with the dataset used to originally develop it The dataset file table must be in its original location must be named the same and must contain the same variables The program will automatically open to the screen where the original analysis was saved OPEN DATASET SELECT DATASET TO OPEN Clicking this button opens a pop up window Select Dataset to Open where the selection of a dataset to be analyzed is made This pop up window is typically divided into several sections Rapid I nsight An
26. c predictive models Such models are used to evaluate new data and apply a probability scores to each record predicting the likelihood of a specific event or condition to occur This product can be used alone or in conjunction with Veera Rapid Insight s data processing and reporting tool This Document This User Manual is intended to provide in depth descriptions of all aspects of the program s function and features Hardware Requirements Systems intending to host Analytics must meet the following hardware requirements Minimum System Requirements e Windows XP SP3 or later e Framework 4 0 and its requirements e 2 GHz 32 bit x86 or 64 bit x64 processor e 1 GB of system memory e 20 MB available hard drive space e 32 MB of graphics memory Recommended System Requirements e Windows 7 x64 e Framework 4 0 and its requirements e 2 8 GHz 64 bit x64 processor e 4 GB of system memory e 1GB available hard drive space e 512 MB of graphics memory Additional Resources For more specialized information try one of these other Analytics documentation resources e Installation Guide provides step by step instructions for installing the program e Getting Started Guide a quick reference and tutorial includes several How To examples e n program Help accessed from the Analytics toolbar menu 1 21 Rapid I nsight Analytics User Manual PROGRAM ELEMENTS Analytics is actually comprised of t
27. e drop down menu gives the ability to change the p value according to how statistically significant the relationship between the X and Y variables should be The lower the zs p value the more significant the relation 0005 DO is ij SAMPLING The Sampling slider adjusts how many records to use to create the model and M M M how many to hold out from development to evaluate a completed model s accuracy Records are chosen at random Current Modeling Sample 100 Current Hald aut Sample 0 MODELING OPTIONS Modeling options control how a model is built and what is displayed in the outcome Use Forward Backward Elimination By checking Use forward backward elimination Analytics will check to see if any of the variables should be dropped from the model during every model building step If a variable that has already entered the model is no longer statistically significant at the chosen p value that variable will be removed from the model Use Fast Model Method Checking the Use fast model method box speeds up model development time by eliminating any variables once their significance level falls below the chosen threshold All variables will be considered during each step if this option is not checked Automatically developed models may also take significantly longer to build on large datasets Rapid Insight Analytics Us
28. e data in report form by clicking Frequency Report Male 1512 60 0 orasan Excel spreadsheet by clicking Export to Excel TWO WAY CROSS FREQUENCIES Crosstab Ey Variable By Value Crosstab Results Gender by Athlete Flag x Haw Humbers Female Athlete Flag 1326 2204 1 13 184 315 Total 1007 1512 2513 How Percentages Column Percentages Total Percentages Two Way Cross Frequencies further breaks down variable frequencies for one or more variables by including another variable First select one or more variables as your Selected Variables from the Categorical and Binary Variables window To see the selected variables by another variable select a By variable from the Categorical and Binary Variables window Click the Submit button to view the frequencies in the window below the variables lists There are four view for each comparison of two variables raw numbers row percentages column percentages and total percentages Users may also choose to view the data in report form by clicking Frequency Report or as an Excel spreadsheet by clicking Export to Excel Means Analysis GENERAL DESCRIPTION The Means Analysis tab performs means analysis on any continuous variable which includes a mean count of the observations max and min for each category or subclass The means for each category of Rapid Insight Analytics User Manual a categorical or binary
29. e the current report Print Clicking Print will print the current report Cut To cut select the text or object you d like to move and click Cut Copy To copy select the text or object you d like copied and click Copy Paste Click Paste to paste the last text or object you have cut or copied Bold Highlighting text and selecting the Bold button will embolden that text Italics Highlighting text and selecting the Italics button will italicize that text Underline Highlighting text and selecting the Underline button will underline that text Left Center Right Justified To change the justification or alignment of text or objects highlight them and click on the justification you d like Rapid Insight Analytics User Manual Font Controls Change the font color size and type by using the drop down menus Preview any combination of these options by clicking the button next to the drop down menus and selecting the color size and type Report Bar Clicking the Report Bar button allows you to drag and drop charts graphs etc from the Report Bar onto a report Notes GENERAL DESCRIPTION The Notes tab creates and saves notes Users can access them via this tab at any time To create a note first fill in the Short Desc field Add more information via the Note window if desired The program automatically numbers each note click on each note number to view the
30. ed as a spreadsheet to Excel by clicking the Export to Excel button below the list of variables Export ta Excel Analyze Data GENERAL DESCRIPTION Use the Analyze Data tab to sort search find sums and means by both decile and category as well as view of the entire dataset Rapid Insight Analytics User Manual ANALYZE DATA RECORDS GRID The data records grid contains all of our observations for each variable ready to be sorted A Rapid Insight Analytics Current Students csv An A File Options Reports Window Help CONTROL PANEL V Deme Decte Select Variable s VARIABLE v Athlete Flag Obs Student ID Gender SAT Math State Athlete Flag Legacy Indicator HS_GPA Scholarship Flag Campus Visit Flag Campus Housing Flag Days betwee STATISTICS MI Campus Housing Flag E zx Male 550 NH 0 Legacy 2 500 1 1 0 3 MI Campus Visit Flag TEE Female 420 NH 0 Non egacy 2 150 1 0 0 m f M M Days between app date and t EES Female 640 M 0 Non egacy 2 550 1 0 0 Gender REELED Female 670 MA 0 Nondegacy 2 100 1 1 0 DATA MIHS_GPA JEZA Female 620 NH 0 Non legacy 2 950 1 D 0 vi Legacy Indicator SEE Female 660 MA 0 Nondegacy 2 350 1 0 0 WISAT Math N E Male 480 MA 0 Legacy 1 550 1 1 0 Ml Scholarship Flag MESSER Female 660 ON 1 Non egacy 2 250 1 0 0 v State Female 530 M 1 Non legacy 2
31. elated to Y NE I Campus isit Flag binary n a no mizsings Related to Y EM Gender categorical n a no missing Related to Y EN H5 GPA continuous no mizsings aM Related to Y I Legacy Indicator categorical n a no missing Related to Y EE I SAT Math continuous no missings aM Related to Y Scholarship Flag binary n a no missings aM Related to Y mH i State categorical n a default Related to Y Student ID continuous no missings aM Related to Y EE VARIABLE CONTROL PANEL TABLE The Variable Control Panel table contains information about each of the variables used in analysis Variable Column The variable column names each variable to be used in analysis Type Column The type column describes each variable s type or distribution If Analytics incorrectly determines the variable type the user can change the type by choosing the best option from the drop down menu Outlier Cap Column Cap Data Value when Often in a large dataset Days between app date and term start o Laus between app date and term start gt fs there are variables that contain one or more atypical values or infrequent observations To ensure the quality of data used in analysis users may wish to set a cap which each value of a given Need ta enter a value Rapid Insight Analytics User Manual variable cannot exceed or dip below To do this si
32. en our X f variables and Y variable should be The lower the p value the more significant the relation is AUTO PROFILE Clicking the Auto Profile button generates a chart that compares the categories of the binary variable chosen to variables for which the two categories have statistically significantly different means or proportions The upper portion of the chart compares each category to each of the continuous variables in a dataset by showing the means of each variable for both categories of the binary variable The lower portion of the chart compares each category to each of the binary and categorical variables in the dataset for which the two categories have statistically significantly different means or proportions By checking the Show Z Score button a Z score measure of how many standard deviations an observation is above or below a mean is included The larger the Z score the further away each observation is from the mean VIEW PROFILE GRAPHICALLY Clicking the View Profile Graphically button displays graphs for each View Profile Graphically variable via the Profile Against Variable drop down menu Options to 66 the lift index add count labels angle the X axis and Y axis labels and print each chart individually are also available Users can display all values by means proportions percentages of sample or both Users can also click the View Chart Data button to view each chart
33. er Manual Show Standard Estimates By checking the Show r Modeling Options standardized estimates box the output will include a column labeled Std Est This column represents the estimates of the regression coefficients from the Coef column standardized to a normal distribution This option is only available when working with an OLS Regression model NSIT se weighted regression se forward backward elimination ast model method Select weight variable Show Multicollinearity Diagnostics Checking the Show Multicollinearity diagnostics box displays any highly correlated relationships between predictor variables Although correlated variables can exist in a successful predictive model the values assigned to their coefficients may not accurately estimate each collinear variable s impact on the Y variable if the other correlated variables were to be removed Diagnostics we can include in our output are Prior Abair e Variance Inflation Factor amp Tolerance The Variance Inflation Factor VIF is equal to the reciprocal of tolerance In general the VIF represents the severity of Multicollinearity by measuring how much the variance of a coefficient is increased because of collinearity e Covariance of the Estimates The matrix for the covariance of the estimates can be found in the Diagnostics section of the output In general a large positive value of covariance can mean that tw
34. gories With field Simply enter the number of records a category must represent to be included in each graph and click Submit This E values option is available if the X axis variable is categorical or binary Granulanty Granularity Slider Rapid Insight Analytics User Manual Use the Granularity Slider to increase or decrease the granularity of the graph Show Transformation Select a transformation under the Show Transformation drop down menu to see if data is improved by applying a transformation to it show Transformation na transformation na transfarmatian laq natural log square root amp quare cubed raat cube Means Proportions 6 of Sample Both The Means Proportions button shows the means or proportions of each X axis variable as Y axis The 6 of Sample button shows the percentage of Means Froportions of Sample the sample each category represents as Y axis The Both button shows the means or proportions as shown with the Means Proportions button as well E S t as a blue line representing the percent of sample that fell into each category These options are available for a categorical or binary variable X variable View Chart Data Click the View Chart Data button to see the sum mean min max and count for each point or bar used in the chart We are able to save this as an Excel file by clicking Save View Chart Data VARI
35. le will save Note that if there are any values which fall outside of the categories created they will automatically be binned together into the category other e For categorical variables name the first category and define which categories fall into the new category selecting is or is not from the drop down menu and then the categories to bin according to this guideline Click validate to create a new category and repeat as necessary When finished creating new categories name the variable in the New Variable Name field and save by clicking Create this Variable Note that if there are any categories not included in the new categories the data contained in them will be binned together into the category other MULTIPLE VARIABLE OPERATIONS The multiple variable operations button allows Lamps Housrxg Fo T TIN 11 Va users to create new variables by performing cn oos MINE EE MEAS Schalasstep Fsg arithmetic operations and or using logic denti statements to combine existing variables Select all of the variables to work with from the Select Variable to Transform window After selecting variables choose to create either a continuous or binary variable The variables selected are listed in the Variable column The Formula Tag column assigns each variable a letter to be used in the Enter a Formula window Enter any operations or logic statements desired to create a new variable When finished click Validate
36. mply click the Set button in the next column and enter the range of values that is appropriate for each variable When satisfied with the settings click Set Outlier Cap to save and exit back to the Control Panel Missing Handling Column Missing Handling identifies missing values within each variable If missing values are found several options are available from the drop down menu e default which can be defined on the Model tab e Fill with Zero e Fill with Variable Mean or e Exclude Entire Record It is important to keep in mind that excluding multiple records can greatly reduce your sample size for analysis do so with caution Model Availability Column A variable s Model Availability column determines whether a column will or will not appear in any predictive model formula Users may select an option from the drop down list e Exclude a variable even if it is related to the Y variable e Make the variable Available even if it is not related to the Y variable and e Allow the software to include the Select Data Record when iiWisia Campus Housing Flag L D variable only If Related to Y D If the variable is data types text date or constant the column will be marked as n a and it won t be available for use in a model Filtering Column Discard Data Record when otherwise The filtering column selects or discards Current Filter Definition records for analysis based on u
37. n be reached by clicking the View Relationships button on the Automated Mining tab This window allows us to view the relationships between our Y variable and each statistically related X variable as a graph VARIABLE RELATIONSHIPS TOOLBAR CONTROLS This toolbar modifies the charts that the Variable Relationships window creates Select and sort which variables to compare and then add labels to charts to make them easier to read changes their formatting to make them more aesthetically pleasing and print them Users can also exclude categories and view these charts by their means or the percentage of the sample that they represent Data records represented in each chart may also be viewed Rapid Insight Analytics User Manual Use Lift Index on Y Axis To view the lift index for each entry on the Y axis rather than the Y value click the Use Lift Index button The lift index shows the Y axis indexed to one and for any value that falls above or below 1 0 we know that the variable mean for that combination of data is above or below the population mean Show Values Selecting the Show Values button labels each bar or point with its Y value 3 D Chart View Clicking the 3 D Toggle Button allows users to switch between 2 D and a rotating 3 D view of a chart This button is available only when the X axis and View By variables are both either binary or categorical To stop the auto rotation click anywhere on the cha
38. o analyze will be selected randomly a Randomly order rows otherwise the will selected in the order encountered in the file table Variables to Exclude from Loading This field is only active prior to clicking the Load Variables button Variables that the user has determined should not be part of the analysis are selected from the list on the left and moved to this field using the Exclude button Variables may be returned to the list of loadable columns by selecting them and clicking the Include button Mark Variables as Identifiers This field is only active prior to clicking the Load Variables button Variables that serve as unique record identifiers in the dataset should not be part of the analysis To keep them from loading the user selects them from the list on the left and moves them to this field using the Identifier button Variables may be returned to the list of loadable columns by selecting them and clicking the Other button Status This field reports the record count and number of variables found in the original dataset Load Variables Button This button temporarily replaces the Load Saved Analysis and Dataset option Clicking the button loads the records into memory LOAD SAVED ANALYSIS ON CURRENT DATASET Users selecting this option will identify a saved Rapid Insight analysis file Load Saved Analysis on ria and superimpose it on the current dataset The dataset file table Current Dataset should
39. o variables are linearly related a negative value can indicate that the two variables tend to move in opposite directions e Correlation of the Estimates The correlation matrix can be found in the Diagnostics section of the output The correlation between two estimates is equal to the slope of the regression line between the variables when they have both been normalized Thus a value of one indicates a direct correlation and the closer to 1 or 1 the value is the more correlated the two variables are The sign of the correlation represents a positive or inverse relationship e Correlation of the Variables The correlation matrix can be found in the Diagnostics section of the output The correlation between two variables is equal to the slope of the regression line between the two variables Thus a value of one indicates a direct correlation and the closer to 1 or 1 the value is the more correlated the two variables are The sign of the correlation represents a positive or inverse relationship e x x Matrix The x x matrix also called the variance covariance matrix can be found in the Diagnostics section of the output This matrix displays the variances of each variable on the diagonal and the covariances between variables as the off diagonal values Use Weighted Regression By checking the Use weighted regression box and filling in the Select Weight Variable field with a continuous variable we opt to assign a weight to each
40. oving variables or use the same method to change a Build Model Automatically model Clicking Run Regression will then regressively analyze the customizations and produce the model Suggest Next Variable Clicking the Suggest Next Variable button allows Analytics to decide which variable from the Variable List window would best improve the model and moves it automatically to the Include these window Explain Model Graphically To view graphs of the relationships between the X and Y variables click the Explain Model Graphically button This action opens the Variable Relationships window see the section Variable Relationships Stepwise Rapid Insight Analytics User Manual By clicking the Stepwise button Rapid Insight Analytics will check to see if any of Stepwise the variables should be dropped from the model following each step of the regression analysis If a variable that has already entered the model is no longer statistically significant at the chosen p value that variable will be removed from the model Exit Regression Return to the main program screen by clicking Exit Regression or closing the window VARIABLE LIST The Variable List window contains all of the variables available when building a model Expand the list of available options by checking the View new variable suggestions box MEMORIZE MODELS To capture a model enter a name in the Model
41. put decile information will allow the program to display a column in your output next to the scoring column that displays the decile or percentile that each observation falls under Start Scoring When you are ready to score click Start Scoring to view a list of all variables in the dataset for selection When you have selected the variables you would like to include click Ok Contact Us For any issues questions about Rapid Insight Analytics please send an email to RI Support at support rapidinsightinc com
42. rger the standard deviation is relative to the mean Missing This number denotes the number of missing null observations for each variable Rapid Insight Analytics User Manual Obs total of observations This is the total number of non null observations for each variable Range The range represents the difference between the highest and lowest observation for each variable Var_type variable type This represents the type of variable continuous binary text etc Distinct categories This represents the number of unique categories within each variable Text This is a yes no indicator for the presence of a text variable VARIABLE STATISTICS LEFT SIDEBAR The left sidebar allows you to see summary statistics only for the selected variables The statistics table adds or removes variables from view making it easy to place variables next to each other to see how they compare Select All Clear All Select Variables Select All displays all of the variables and their statistics at once Clear All clears all variables from the table while Select Variables individually displays variables by clicking the checkbox next to each one Variables Statistics Report A report of the variable statistics selected can be generated and saved exported by clicking on the Variable Statistics Report button below the list of variables Variable Statistics Report Export to Excel The statistics can be export
43. riable sub string for text variables and pai root days from for date variables After defining a new variable name it in cubed root the New Variable Name field and click Create this Variable It will then show up in the New Variables Created window Transformation CREATE A BINARY VARIABLE Binary Variable 1 when Select Create a binary variable and choose is mm CMale a variable The options will be different depending on the data type of the variable selected Generally users will be prompted to choose a category or an operation lt to Binary Yariable 0 when otherwise Rapid Insight Analytics User Manual etc to sort values between the 0 and 1 categories BINNING To create a categorical variable from a continuous or another categorical variable select Binning and a variable Depending on the data type of the variable selected the next steps differ e Binning binary variables doesn t usually make sense but one helpful feature of binning is the Is Missing checkbox which allows assigning a binary value to any missing data cells e For continuous variables users are prompted to name our first category and define which values fall into this category To move to the next category click Validate and repeat each step Be sure to name the variable in the New Variable Name field When finished creating categories click Create this Variable and the variab
44. row equal to the value of the weight variable for Rapid Insight Analytics User Manual that row A weight value of zero would therefore cause a row to be ignored and a weight value of two would cause that row to be weighted twice as though that row appeared twice in the dataset Static Creation of Automine Variables This option is currently not available BEGIN REGRESSION ANALYSIS When the regression type and options have been chosen click Begin Regression Analysis The program will build a matrix of variables in preparation for the regression analysis Building a Model to Predict Y Variable GENERAL DESCRIPTION This screen is where models to predict the Y variable are built Rapid Insight Analytics can build a model automatically or users can manually create their own custom models REGRESSION OPTIONS By using the Regression Options users can customize their regression model Build Model Automatically The easiest and fastest way to build a model is by clicking the Build Model Build Model Automatically button This option allows Rapid Insight Analytics to build what it feels is the best model that can be built on the current dataset It takes all non linearities in the data into account and makes all of the appropriate transformations on the candidate variables automatically Automatically Run Regression Users may choose to build their own models by manually adding and or rem
45. rt and the chart will freeze To see other views drag the scroll bars on below and to the right of the chart Sort Bars by Height The Sort Bars by Height button is available if the X variable is binary or categorical This button orders the bars from largest to smallest left to right Angle X Axis Y Axis Annotations These options angle the X axis and Y axis labels to improve legibility Show Regression Clicking the Show Regression button superimposes a regression line fitted to the variable chosen A regression line is only available when our X variable is continuous Print Current Chart EF Print the chart in view by clicking the Print Current Chart button Related Variables The Related Variables drop down menu contains only the variables that Related Variables Rapid Insight Analytics chose as having a statistically significant HS GPA relationship with the Y variable during the Automine process View By Select a variable to View By to view it along with an X axis and Y axis View Bu Genet variable all on the same graph If the X axis variable and View By variable Gender I are both either binary or categorical the resulting chart will be 3 D To switch to a 2 D view use the 3 D toggle button on the upper toolbar Exclude Categories With Exclude those categories that don t have enough values for analysis by filling in Exclude Categories with the Exclude Cate
46. rts Variables button will create a Create Report On These Variables series of charts for the relationship between each predictor variable and the selected Y variable that can be printed or saved as needed Export Menu Create PowerPoint Slide Show on these Variables This button creates a PowerPoint slideshow containing one chart per slide Create PowerPoint Slideshow On These Variables comparing each predictor variable to the selected Y variable T Sau eRe HI i E H s x E Palod Varnshle n ranskan Create Variables GENERAL DESCRIPTION The Create Variables tab allows users to create new variables from existing variables Choose what type of variable to create and rename variables in this tab Once a transformation type is selected only the variables that can be transformed in this way will be shown in the Select Variable to Transform window Each created variable will be available in all other areas of the application such as the Control Panel and Univariate Analysis tabs We can view any created variable s definition in the Current Variable Definition box by highlighting that variable LUTTE TN SINGLE VARIABLE OPERATIONS After clicking the Single Variable Operations button and choosing which variable to transform options for transformation from the Transformation drop down menu appear such as taking the log or square root of a continuous va
47. s click the Export Analysis button under the File menu Here you will be prompted to select which variables you d like to export and name the file you d like to save the analysis to Close Analysis To close a single analysis at any time click the Close Analysis button from the File menu You will be prompted to save any unsaved analyses at this time Close AII To close all running analyses at any time click the Close All button from the File menu You will be prompted to save any unsaved analyses at this time New Analysis To perform a new analysis on the current dataset click the New Analysis button from the File menu To toggle between analyses we select the one we wish to view from the Win Exit To exit the Analytics program at any time click the Exit button from the File menu or the You will be prompted to save any unsaved analyses at this time Rapid Insight Analytics User Manual OPTIONS MENU Options Reports Window Help Novice Green Mode v Novice Green Mode The Novice Green Mode applies to color green to the button that is the next logical j step for the user and is a good way to be Dataset Evaluation Count P guided through the process of analysis Dataset Query Timeout d Color Palette v Load new Dataset columns inta Analysis If you wish to change the color palette of Analytics you may do so by selecting a new palette from the Color Palette drop down menu
48. s data PROFILE REPORT Clicking the Profile Report button automatically generates a report of the values shown in the Auto Profile chart In this window we can print Rapid Insight Analytics User Manual the report or send a copy of it to a word processing program EXPORT TO EXCEL These profiles can be exported as a spreadsheet to Excel by clicking the Export to Excel button below the list of variables Model GENERAL DESCRIPTION Regression models are powerful tools for predicting the likely response of one variable Y variable based upon on the known behavior of other variables The Model tab allows users to set specifications about which variable to model on and how accurate it should be REGRESSION TYPE r Regression Type Analytics uses regression analysis to g develop its predictive models OLS Regression Regression uses several techniques for modeling and analyzing data describing past behavior More specifically regression analysis helps explain how the typical value of the Y variable changes when any one of the other independent variables is changed while the other independent variables are held fixed This is known as a step These series of stepwise results are then combined into a formulaic whole the predictive model Analytics offers two types of regression analysis Logistic and OLS Ordinary Least Squares Logistic Regression Logistic Regression
49. selecting the variable from the list Distribution of Athlete Flag ANALYSIS VARIABLE 5 CS Dae between app date and ender LIST Sct HS_GPA nnn E Legacy Indicator OAT Math ANALYSIS Scholarship Flaq n To perform a gt MULT IVAR l A mE univariate analyses ANALYSIS f UN AUTOMATED Pie Bar or a specific MINING variable j select it rei S View Percentages from the variable FREQUENC list MEANS ANALYSIS UNIVARIATE mm ANALYSIS CHART Print Chart A R E A Show Missing Values The chart area REPORTS Tis ical NES displays one chart at a time for each selected variable and chart type Pie Bar chart types If a variable is a categorical or binary variable its distribution will automatically be shown as a pie chart If the variable is continuous its distribution will be shown as a bar chart Users can choose to select the preferred chart type View Percentages View Counts Add the counts or percentages for each distribution by clicking the View Counts or View Percentages buttons View Counts View Percentages Rapid Insight Analytics User Manual View More Bins Choose View More Bins to spread data out into more categories to see View More Binz the effect that a different categorization setup has on each variable Find Outliers If a continuous variable has values that are atypical outliers click on the Find Outliers
50. ser entered E criteria To filter select a variable and click the Set button at the end of the row Use drop down boxes in the Filter Dataset on Variable window to define the filtering criteria Click Set Filter to save the filter and exit back to Control Panel Y VARIABLE Select one variable to be the variable to be modeled the Y Variable by either highlighting it then clicking the green downward arrow or by double clicking the variable Only binary and continuous variables may be selected EXPORT TO EXCEL Save a copy of the Control Panel to Excel by clicking the Export to Excel button Export to Excel Rapid Insight Analytics User Manual Variable Statistics GENERAL DESCRIPTION The Variable Statistics tab allows users to view simple statistics for each of the variables giving a clear and concise way of gaining insight into the dataset It is a good idea to check these statistics because problems such as data omissions and outlying points are often detected here ae er ar s g LS A File Options Reports Window Help CONTROL variable mean std dey min max coeff var missing E obs range var_type E distinct text PANEL o Athlete Flag 0 1250 n a 0 1 000 1 000 binary Campus Housing Flag 0 1520 n a 0 1 000 1 000 binary Select Variable s Campus Visit Flag D 4002 n a 0 1 000 1 000 binary VARIABLE M Athlete Flag Days between app date and term start 204 18 100 3
51. ssesseeeceeseess 15 Analyze D I M EPUM MP NM 15 EINER 15 ANALYZE DATA RECORDS GRID erur E EEEE E E E EAE E A 16 ANALYZE DATA 16 UAN aa POV GW SIS rE E N E E E E 17 Rapid I nsight Analytics User Manual GENERAL DESCRIPTION eeseeseeeeeeenn enne nnne nnne nnn nnne nne nnn nnn rentes ease esee sre ses eas ee sessi en nis 17 UNIVARIATE ANALYSIS VARIABLE LIST E snes NK TUR MARIS SuSE 17 17 IER 18 GENERAL DESCRIPTION cccccccecceccsecseceeeeecceeeeeeeeeeeeeeeeeeeeseesesaeeeeeeeeeeeseeeseeaeseesseceecseeseeeeeaeceeeaeenees 18 MULTIVARIATE UPPER TOOLBAR CONTRO LS pU 19 MULTIVARIATE LOWER TOOLBAR CONTROLS escuckosesc vh eR E Kk xa E REESE Eae eo pereo 20 Automated cies E ecu Us erae CE PRU 21 No c 21 ANALYSIS VARIABLES R
52. to the Y axis variable via the ans Variable drop down menu If the X axis variable is continuous the output will be in the form of a point graph If it is categorical or binary the output will be in the form of a bar graph SAT Math View By Select a variable to View By to view it along with an X axis and Y axis variable all on the same graph If the X axis variable and View By View By variable are both either binary or categorical the resulting chart will be Gender 3 D To switch to a 2 D view use the 3 D toggle button on the upper toolbar Exclude Categories With Rapid Insight Analytics User Manual Exclude those categories that don t have enough values for analysis by filling in the Exclude Categories With field Simply enter the number of records a Exclude L ategories with category must represent to be included in each graph and click Submit This Mu values option is available if the X axis variable is categorical or binary Granularity Granularity Slider Use the Granularity Slider to increase or decrease the granularity of the graph Show Transformation Select a transformation under the Show Transformation drop down menu to see if data is improved by applying a transformation to it show Transformation na transformation na transfarmatian laq The Means Proportions button shows the means or proportions of each X axis natural log
53. ual RELATED VARIABLES The Automine process automatically separates the variables according to whether they are statistically related to the selected Y variable Related variables appear in this column Note that Automine runs independently of the Control Panel For example variables designated as Exclude in the Control Panel may still appear in the Related Variables column NON RELATED VARIABLES The Automine process automatically separates the analysis variables according to whether they are Statistically related to the Y variable or not Non related variables appear in this column CHANGE P VALUE The Change P Value drop down menu gives the ability to change our p value according to how statistically significant we mnm require the relationship between the X variables and the Y variable to be The lower the p value the more significant the relation is Change p value AUTOMINE IT To begin the Automine process simply press the Automine It button Each analysis variable is tested against the Y variable to determine if a statistically significant relationship exists between the two variables Autamine It VIEW RELATIONSHIPS The View Relationships button becomes available after the data is automined It allows users to visually compare each related variable to View Relationships the Y variable Variable Relationships GENERAL DESCRIPTION The Variable Relationships window ca
54. uring our analysis When we have secured all of the graphs and charts we want we can export them all to a report or a PowerPoint presentation by selecting either toggle button on the bottom of the report bar and clicking Export Clicking Clear All will erase all collected diagrams from the report bar Variable Statistics Clicking the Variable Statistics auto generates a report that includes the number of dataset and sample records dataset variables analysis variables and the statistics that go along with each analysis variable including mean minimum and maximum values variable type number of missing values Rapid Insight Analytics User Manual standard deviation number of distinct values and any filters or caps you may have applied during analysis Control Panel GENERAL DESCRIPTION The control panel tab provides us with the tools to make decisions about our data based on our own criteria For example if we want to examine a specific period of time we need to filter out records that do not fall within the given time span We also have the ability to change variable types and model availability here Importantly the Control Panel is the tab that is used to designate and edit your Y variable Variable Control Panel Variable Type Outlier Cap Missing Handling Model Availability Filtering E D Athlete Flag binary n a no missings If Related to Y Campus Housing Flag binary n a no missing R
55. variable can also be further sub classed by a second categorical or binary variable To see a report of this data select the Means Report button or to view as an Excel spreadsheet select the Export to Excel button Variables to Examine Means From the variable list select a variable to examine the mean of by highlighting the variable and clicking the green arrow to move it to the Variables to examine means of Varables to examine means of SAT Math window You may select more than one variable to examine the means of in this step The means that appear in the output table are the means categorized by any other variables selected for the variables selected in this field Take Means By m From the variable list select a variable by Bl Include Missing as category highlighting it and clicking the green arrow to move it to the Take means by window Next click Submit if you are ready to view the means analysis Sub Class By z Further subclass the means analysis of variables sub class by Include Missings as category by selecting a third variable to View By Do this by highlighting a variable from the variable list and clicking the green arrow to move it to the Sub class by window Next click Submit when ready to view the means analysis Correlation Analysis GENERAL DESCRIPTION The Correlation Analysis tab displays positive and negative correlations between variables The SAT
56. ve percentage increase in accurate prediction for the given decile and lower Rapid Insight Analytics User Manual Compare Models GENERAL DESCRIPTION The Compare Models tab provides the ability to evaluate the models memorized on the Model tab by comparing them to each other Click on each model to view them individually or select multiple models to compare by holding down the CTRL or SHIFT key and clicking on each model to compare Control what to view about each model by checking the boxes Model Data Decile Analysis and Cumulative Lift on the right side of the screen To display labels on a graph simply click on it Graphs are viewable by decile twentile and percentile by selecting an option from the drop down menu Model Report Click Model Report to generate a report that can be printed or exported as a text document Export Model to Excel These model comparisons can be exported as a spreadsheet to Excel by clicking the Export Model to Excel button below the Model Report button Saving Scoring Model In order to score a model later one save it here by selecting just one model from the model diagnostics table and clicking the Save Scoring Model button A file with an rism file type is produced What If GENERAL DESCRIPTION The What If tab allows users to explore how the Y variable would be affected if one or more of the related variables were increased or decre
57. wo standard and one optional modules that appear under the PROGRAMS RAPID INSIGHT INC ANALYTICS section of the Windows START menu Though all of these are typically installed licensing determines which are made available to a specific user Analytics Module The module labeled ANALYTICS is the core of the program This is where all aspects of data analysis predictive model building and reporting are accomplished Scoring Module Scoring is a module that is available to all standard licensed users It applies a predictive model developed with the Analytics program to new data This module assigns a probability score to each record based upon its unique combination of values VRE Variable Reduction Engine Module The Variable Reduction Engine VRE is an optional module whose purpose is to work around limits imposed by ODBC and OLEDB drivers on some file based datasets VRE works in conjunction with the main Analytics program to permit the use of datasets with more than 255 variables Rapid I nsight Analytics User Manual ANALYTICS MODULE Select and Load a Dataset GENERAL DESCRIPTION This is always the first window encountered when opening the Analytics module This is where the dataset to be analyzed is selected It is also where any stored analysis may be recalled and applied to the dataset DATA SELECTION CONSIDERATIONS Analytics always performs its analysis and model building upon a single dataset This dataset may
Download Pdf Manuals
Related Search