Home

User Guide - SAP Service Marketplace

1. Time stamped Population Editor 5 Fields A Merge Y Fitters 8 Prompts view Data E SQL Statement E Documentation t rF O vv amp pommel DateTime Functor Create a new field as an aggregate tege Sti The panel Define an Aggregate is displayed 3 Inthe section Event Tables Selection of the tab Aggregation Settings use the Table list to select the table that contains the event data you want to aggregate Define an Aggregate Aggregation Settings Period Settings Filters and Pivots Settings Not compatible with Average Number of fields that wil be created 0 oe as Cate rt ctmnatisece CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 33 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 4 If needed enter an alias for the table in the field Alias 5 In the list Date Column select the column that contains the event dates Define an Aggregate Aggregation Settings Period Settings Filters and Pivots Settings Reference Table Key Events Table Key id E Aggregate Operation Specification Target Column account id Not compatible with Average Not compatible with Average ZZ In the section Join Keys select the reference table field to use for the join in the list Reference Table Key 7 Select the events table field to use for the join in the l
2. 1 In SAP Infinitelnsight start panel select the feature Perform a Data Transfer in the Toolkit section SAP InfiniteInsight Recommendation A Toolkit Open the Data Viewer Perform a Data Transfe List Distinct Values in a Transfer a data source to another location or into another format Get Descriptive Statistics for a Data Set The panel Select Data Set Source is displayed 2 Select aData Type aFolder CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 55 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer anda Data Set which is the entity you want to transfer into a new database it is preceded by the icon with a green arrow in the middle Select Data Set Source Data Type Date Base Data Selection x Folder Ksrvmulti2_sqserverz005 ili E Browse Select a Data Set Data Set Entity 3 F a Browse Metadata Metadata are stored in the same di Analytical Record E i Time stamped Population H Performance Indicator E SAP _ALL I304809 ModelToDelAlain 0 SAP_ALL I304309 ToDelAlain A AD_DW Adventure WorksDWBuildVersion AD_DW DimAccount gt AD _DW DimCurrency E AD_DW DimCustomer A AD_DW DimEmployee 0 AD_DW DimGeography 0 AD_DW DimOrganization A AD_DW DimProduct 0 AD_DW DimProductCategory FA an Mi aeai haama 3 ClickOK Click Next 5 Inthe pop up window Transfer a Data Mani
3. Based on Entity entity_disposition Timestamp KxTimeStame l 7 Edit Filters Change KxTimeStamp Storage Fe os 3 Inthe Name field enter the name of the time stamped population you want to create 4 Click the button Edit Filters The Time stamped Population Editor opens Time stamped Population Editor CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 29 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Editing the Time Stamp Prompt Scenario The time stamp created in the default time stamped population is a prompt for which the default value is the date of its creation You can choose to modify this prompt to better fit your needs Settings to Use Default Value enter the date 1998 01 01 12 00 00 Prompt Heading do not change the default value vV To Edit the Time Stamp Prompt 1 Inthe Time stamped Population Editor click the Prompts tab Time stamped Population Editor Si Fields dy Merge Y Fiters 10 Prompts view Data T SQL Statement E Documentation analytical Data Set Reference Date Prompt Editor SX Name TimeStampPrompt Type DateTime E Default Value 2014 02 05 10 00 00 aa S Prompt Heading Analytical Data Set Reference Date See Example CAE 4 Inthe field Default Value enter the value you want fe Prompt Editor CUSTOMER SAP Infi
4. CUSTOMER End User Documentation Document Version 1 0 2014 11 A aa Se Y AS 7 we s k x N i X h NS AN l f 4 ee W l U NS 4 a i Wi A r a ZON Nw j N Sgi k 1 ae 3 1 32 Jo 3 4 3 5 4 1 4 2 Welcome to this GUIS ccavaccasacauiecaciascaassuvaducacsuiecusussuveaccdussseaunsuensacdedsasequusiscatdusuaaaeduvadasdedtvansuvddeuatanineuneecuveaae 3 WV Feats Document COW CIS aumeneneninr a e as e E N 3 Essential ONCEIILS mii A A ACA Aa Aa AAAA S A a 4 DEMU ON ties he aati a a a a a a aa 4 2 1 1 Infinitelnsight Explorer Semantic Layer s n nnnnnssssssssernrrrnrirrrrrrrrrsssettriirrrrrrrrrreesrrrtrrrrrrrrrrrnrrrns 4 22 ERU ana a aN E E EAA A A A nly aa A eddie 4 2 1 3 Tmesstamped Foppa oane r eerde ien deden 5 2 1 4 ARIAIVEICAl RECO O a tal tele tal tile tate las ade dk de he ahaa dell 5 ZA Anaea DaI OCU oea ston uecuepnonte ca cotarts saceooss Saiasoas O 5 2 1 6 Tempor Amalvtical Date o E Eora E E TE NANNA 5 2 1 7 Perlormance WMG ALOR PE esaa aaar aaar aae 6 MET O COLO GY aneen e E EE EEE E E E a d d Wel ciel d Miele a a a aa 7 Application Scenario Segmented Cross sell in Retail Banking scccsessseeccnnsseeeesensssseeennseseees 8 Present aos menan a aaa a a an dng a te ee ee a Rell ehel neds mudueeduenuancia 8 Creating the Performance Indicator stararstetssncatatcsarstetatarsiadahersd oabaacbaechciebadsbtinbdanh taped ainitetactadhdaceaeGsdadneathcateeataeates
5. Explorer Analytical Data Management v7 0 17 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 3 Preparing the Data 3 3 1Creating the Entity Scenario In this particular case even if the problem statement for the project speaks about selling credit card to customer since credit cards are associated with accounts you have to determine together with everyone involved in this project if the entity for this project is the account on which the credit card is attached the customer the only one that can receive a piece of mail describing the offer or the association between the account and the customer which is called a Disposition in the database schema describing the role of the person with the account There are two roles in our example one is the owner of the account and another possible role is a user of the account To define an entity you must take into account the fact that this entity makes business sense that it can be characterized in terms of attributes that it can be associated with predictive metrics in relation with the tasks you want to perform in this case the number of gold credit cards Defining an entity is not a minor operation Entities may be used in many projects and cannot be changed without an impact analysis on all the deployed processes using this entity The entity selected is the Disposition
6. Previous CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 12 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 8 Inthe list Event Table select the table containing the data on which you want to create the aggregate Shifting Aggregate Aggregation Settings Filters and Pivots Settings Event Table redit_card DateCoumn issued date Ol Aggregate Operation Specification Number of fields that wil be created 1 oe Cancel Previous albis CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 13 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 9 Inthe list Date Column select the column which contains the date the aggregate should be based on 10 Inthe Function list select the aggregation operation you want to apply The following operations are available Functions Description Returned Values Count computes the number of not null values found in the number of not null occurrences in the selected selected column of the Event table column l Note This may differ on some databases and behave as a count no operand Count computes the total number of lines in the Event table total number of occurrences no operand Sum compute the sum sum Average compute the mean mean Min identifies the minimum value minimum value Max identifi
7. Settings to Use Enterentity_disposition asthe Name of the entity Select disposition as the Table Select disp_idas the ld Field Donot define a Filter vV To Create an Entity 1 Inthe menu of the panel Explorer select Entity 7 New Delete Edit A 4 n view Data Switch View Filtering Options Export Analytical Record Create a New Entity Time stamped Population Temporal Analytical Data Set Performance Indicator The panel Edit Entity is displayed CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 18 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 2 Inthe field Name enter the name of the new entity Edit Entity Description w ir nae Id Field m 7 Edit F E You must select the table the entity will be based on 3 Inthe field Description enter a description of the new entity 4 Toselect the table on which you want to build the new entity click the Browse button and select a table in the displayed dialog box Click OK to validate your choice 5 Inthe drop down list d Field select the column containing the entity identifier Edit Entity 6 Optional to add a filter to the entity in order to remove some lines depending on a specific condition click the button Edit Filter The existence of a filter is indicated as listed below T no fi
8. 22 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking VI To Create an Analytical Record 1 Inthe panel Explorer click the menu and select Analytical Record Explorer 012 10 02 11 17 37 2012 10 02 li 17 47 Default Time stamped Population ii Performance Indicator Metadata Metadata are stored in the same place as data source ag Ges The panel Edit Analytical Record is displayed Edit Analytical Record BA Name Description Based on Entity a entity _dsposition version 3 Fi Gs Ch vamutnone nem CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 23 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 2 Inthe Name field enter the name of the new analytical record 3 Inthe Description field enter a description for the new analytical record 4 Inthe drop down list Based on Entity select the entity that will be used to create the new analytical record Edit Analytical Record Based on Entity entity _dspostion Versi on 3 vers 5 Click the button Edit Attributes to add computed fields or fields from other tables to the new analytical record For the detailed procedure see To Add Fields from Other Tables below 6 Optional you can associate the analytical record attributes to specific domains By defa
9. 3 The third step during which you define the time stamped population to be analyzed will be executed by operation or business analysts and can be redone several times to adapt the population to your needs 4 The last step is the automated creation of the temporal analytical data set consisting of the product between the analytical record which determines the number of columns and the time stamped population which determines the number of lines Infinitelnsight also offers the possibility to create a target on the fly when it does not exist in the original data CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 7 2014 SAP SE or an SAP affiliate company All rights reserved Essential Concepts 3 Application Scenario Segmented Cross sell in Retail Banking IN THIS CHAPTER FS Se MAU OI oases ae itera eta E me se eters sete E aren 5 dione lo mn oe Scere eyes ee erase Ue reste se cares eee eee 8 Creating the Performance INdiCator ccccccccseeseeccceeeeeeeseeeeeeeeeeeaeseeceeeeeeseeeeseeeeeeessuseeeeeeeeessaeeeseeeeeesseseaeeseeeeeeaaas 8 FNS AMIN Ne Data reste penetrated sect ie seis wind arene palette a sale te ode nian da pee te wee asia EE 18 Creating the First Classification MOEel cccccccccccceceeeeeeeeeeeeeseeeeseeeeeeeesaeeeeeeeeeeeseeeseeeeeeeessaeseeeeeeeessseaeeseeeeessaas 44 Analyzing the Model IR SUINS csgccscceccntdccecdanesccozeidnensddecasnsdededdesddccedesdoniecashadlosc
10. Data Set field to select the file or table containing the data Select a Data Source Use a Fie or a Database Table Use Explorer Data Type Text Fie Folder aff Samples Census g Browse wast S R Bronce af Cutting Strategy Metadata Metadata are stored in the same place as data source 2 In Use Explorer Use the Browse button corresponding to the field Database Source to select the database containing the analytical record and time stamped population If you want to access a protected database do not forget to enter the username and password when selecting the database Use the Browse button corresponding to the field Analytical Record to select the analytical record you want to use In the window Data Source Selection all the existing Analytical Records are located under the tree item Analytical Record ET x Select an Analytical Record Use the Browse button corresponding to the field Jime stamped Population to select the time stamped population you want to use In the window Data Source Selection all the existing Time stamped Populations are located under the tree item Jime stamped Population Select a Time stamped Population E ga Time stamped Population E ActiveDispositionNoCard Date CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 46 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail
11. Operators m l Categories Miscellaneous Operators J String Operators ied Assocation Associate by Position 2 Messages j expression is correct 6 Click OK 7 Enter the name of the variable in the Name field 4 Enter the Computed Field Name Name creditcard bought Press Ok to Validate 8 Click OK and then Next to define the variable vV To Define the Target 1 Inthe Filtered Time stamped Population panel click the Target tab Filtered Time stamped Population D Debug View Data SOL Expression ActveDisnesitionNoCard Date CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 2 Inthe Target field select the target variable you have just created Filtered Time stamped Population 8 2 g Debug View Data SOL Expression Name ActiveDisposi tonhioCard Gis te Description Default Time stamped Population Settings Target no 3 Click Next 3 3 4 Copying Objects to Another Metadata Repository The button Metadata allows you to specify the location where the metadata should be stored This Copy option can be useful when you are using different metadata depending on the team you are working with Explorer a E E Analyse Record JAR _CientGeoDemo i 0130117 1209 13 Metadata are stored in the same place as data sour
12. Time stamped Population Modifying the Default Time stamped Population Scenario A default time stamped population named entity_dispositionPopulation has been automatically created when you have created the entity_disposition entity You will need to modify this population so that it contains all active dispositions for which no cards have been issued on January 1 1998 An active disposition is a disposition for which at least one transaction has been performed in the 6 months prior to the selected date CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 28 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Editing a Time stamped Population Settings to Use Name ActiveDispositionNoCard Date Based on Entity entity_disposition Time Stamp KxTimeStamp vV ToEdit a Time stamped Population 1 Inthe Explorer panel select the time stamped population you want to edit Explorer op w g E E J debug Toobox k New Delete Edit pann Switch View Filtering Options St qo i iaa OOS SS a ee E E Atty dianosiionPonuat TEF sa z fault Time sts Popula tar AR eset es lt 7 17 12 09 13 Metadata Metadata are stored in the same place as data source Filtered Time stamped Population pa a Debug View Data SQL Expression Name enttity_dispositionPopulation Descripton Default Time stamped Population Settings Target
13. each with its maximum time mmediate duration lt 1s The query is accepted and executed immediately Batched Is lt duration lt 2 s The query is accepted but will be executed on next idle time Rejected 2s lt duration The query will never be executed The number names and limits of classes are defined by the user in order for these values to match the current DBMS configuration and DBMS usage policy CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 47 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking The Explain Mode has been Configured If the Explain mode has been configured by your DBMS administrator there are two possible outcomes to a query the query is accepted and executed this is completely transparent SAP Infinitelnsight accesses the data without further input from the user the query needs to be validated before being executed a pop up window opens displaying a message configured by the DBMS administrator A query that needs validation can be categorized in two ways medium sized You will probably have to check with your administrator to know which action to take Ifthe administrator authorizes the query click the Continue button The pop up window closes and the requested action Is carried out fthe administrator does not authorize the query click the button Stop Query the pop up wi
14. for the target by clicking the icon Edit Field Mapping The panel Edit Field Mapping appears By default there is no mapping defined Data Manipulation Transfer Target Settings Target Settings Target Database Ksrvmulti2 _sqiserver2005 a Browse Transfer a entity disposition O O a aaa oe Browse C Target Metadata Repository Metadata are stored in the same place as data source Browse Table Mapping Target Table Edit Field Mapping Target Field e No mapping was defined for the field disp_id um Guess el Gs CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 57 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer 9 If you have selected a correct Target Table in the previous step click Guess The Target Fields are automatically filled in i Edit Field Mapping 10 Click OK The green arrow on the icon Edit Field Mapping indicates that the mapping is done Data Manipulation Transfer Target Settings Target Database Kervmuli2 sgserver2005 O e rane ae e ree C Target Metadata Repository Metadata are stored in the same place as data source r Browse Table Mapping Edit Field Mapping 11 Click Check A pop window indicates that the mapping for the transfer is correct i i The mapping for data manipulation transfer is correct 12 Click OK 13 Click Next A pop up message indicates that th
15. information If you look at the Contributions by Variables see graphic below you can see that one of the most important variables is the disposition type Maximum Smart Variable Contributions 0 025 0 050 0 075 0 100 0 125 0 150 0 175 0 200 0 000 bith_numbes A2 AB A11 A4 AS A15 Al0_cte A10 SE AB A13_cte Variables district_id Al4 A3 A7 If you click this variable on the graph you will display the categories significance that will show that a customer who is owner of an account is more likely to buy a credit card than one who is only a user of the account Variable disposition_type Influence on Target 0 05 0 00 0 05 pd Th O oy a m iA DIZPONENT E Validation CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 54 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 4 Performing an Entity Transfer It is possible to migrate an Entity and its Time stamped Population while performing a data transfer This option is available via the SAP Infinitelnsight Toolkit feature The data transfer process encompasses various steps starting from the creation of the entity and its related time stamped population in SAP Infinitelnsight Explorer to performing their transfer in the SAP Infinitelnsight Toolkit 4 1 Transfer an Entity vV To Transfer an Entity
16. one of the greatest value added features of SAP Infinitelnsight models the discovery of non intuitive information To create your model you use data where the target is already known however to be able to predict correctly your target when you apply the model you need to exclude from the explanatory variables any information that allowed you to create the target In this scenario it means that you need to remove from the explanatory variables all the variables coming from the table in which the credit card information is stored Settings to Use Target Variable creditcard_bought Weight Variable none Excluded Variables KxIndex_1 card_id card_type issued_date vl To Select the Target Variable 1 Onthe screen Selecting Variables in the section Explanatory Variables Selected left hand side select the variables you want to use as Target Variables By default the last variable of the data set is selected as the target variable l Note On the screen Selecting Variables variables are presented in the same order as that in which they appear in the table of data To sort them alphabetically select the option Alphabetic sort presented beneath each of the variables list 2 Click the button gt located on the left of the screen section Target s Variable s upper right hand side Target Variables creditcard bought C Alphabetic Sort The variable moves to the screen section Target s Variable s Also selec
17. 0 lines of the initial data file In order to avoid all bias we encourage you to mix up your data set before performing this analysis CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 49 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Each variable is described by the fields detailed in the following table The Field Gives information on Name the variable name which cannot be modified Storage the type of values stored in this variable Number the variable contains only computable numbers be careful a telephone number or an account number should not be considered numbers String the variable contains character strings Datetime the variable contains date and time stamps Date the variable contains dates Value the value type of the variable Continuous a numeric variable from which mean variance etc can be computed Nominal categorical variable which is the only possible value for a string Ordinal discrete numeric variable where the relative order is important Key whether this variable is the key variable or identifier for the record O the variable is not an identifier 1 primary identifier 2 secondary identifier Order whether this variable represents a natural order There must be at least one variable set as Order in the Event data source Warning If the data source is a file and the va
18. 8 32l Cea Ne Caed A E a a E A N NA 9 32 2 Creating the Performance Indicator asic ricee atlases accel ntaerd tinted An enend a Dawend a eead seareedetndeerdeeedens 10 FT AMIN Mea as a Sat tc te E T R helen te elas cil lee wie al i ch Nal eR AS ASS ABS 18 331 CC ate TASEEN E e Aa a aa N OAS eat 18 3 92 oreatihe The Altaly UICALINCC OF Gl aA A ENAA 22 3 35 95 Updating the Default Time stamped Population cccccccscccsccceeeccececceeeeeeeeessnsstseeeeeeeeeeeeeeeeesenans 28 3 3 4 Copying Objects to Another Metadata Repositol Vases ea ees eae le eee 42 Creatine the First Classification WOE x scricchcctentlaceiaal ea a tint e a aE A REAT eens 44 3 4 1 Selecting the Data to be Modeled 2 o oo ccceceecsscccceeeeeeeeececeeeeeecesessssaaeeeeeeeeceeeeeeseseseeeetstsssiaaeeeeeeees 45 3 4 2 DescrHDINE TING DIa een aaa ae el bead 49 3 4 3 Selecting the Ve SS aa aaa a a al lear cet AAEE AECE TINSEN maa sun aa airasia 51 3 4 4 Checking the Model Paramete sacs catia caties A A A A 53 Analizine the Model RESU aneian a Seme eee Semele celeesSenclnn gets 54 Periorming n Entity transferon aaa aaaea aa aaa aaa aa Ea aE aa aaa a aaa anaE aana 55 Iransior aN EMY oann aa E UE E E a eae uones 55 Use the Transferred Entity in a New Database incccvicicicvudcicdevidcnceeveeee deed baasenuslesesncceanecevecacececesssivavinsdennddeddeasdes 59 CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 2014 SAP SE or an SAP affiliate company All righ
19. Banking 3 If need be create the target see section Creating the Target on the Fly see Creating the Target on page AO 4 Click the Next button Depending on the size of the query sent to the database an additional warning may be displayed asking you to validate its execution For more information see Special Case Data Stored in Databases the Explain Mode on page 47 5 If the time stamp for the population is a prompt it is displayed asking you to either validate the default date or to enter another date x Reset Analytical Data Set Reference Date 1998 01 01 12 00 00 wait OK Cancel 6 Click the OK button Special Case Data Stored in Databases the Explain Mode Before requesting data stored in a Teradata Oracle or SQLServer 2005 database SAP Infinitelnsight uses a feature called the Explain mode which categorizes the performances of SQL queries in several classes defined by the user In order to be as fast and as light as possible this categorization is done without actually executing the full SQL query l Note 1 For all versions of Teradata 2 For all versions above and including Oracle 10 The objective is to allow you to estimate the workload of the SQL query before executing it and then deciding possibly thanks to an IT Corporate Policy if the SQL query can actually be used For example an IT Corporate Policy may favor interactivity and then define 3 classes of SQL queries
20. able of data To sort them alphabetically select the option Alphabetic sort presented beneath each of the variables list 2 Click the button gt located on the left of the screen section Excluded Variables lower right hand side Excluded Variables is sued_date Number of Variables 25 Alphabetic Sort dh OO Alphabetic Sort The variable moves to the screen section Variables excluded Also select a variable in the screen section Variables excluded and click the button lt to move the variables back to the screen section Explanatory variables selected 3 Click the Next button The screen Parameters of the Model appears CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 52 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 4 4 Checking the Model Parameters The screen Summary of Modeling Parameters allows you to check the modeling parameters just before generating the model Summary of Modeling Parameters Model Name creditcard_bought_ActiveDispositionNoCard Date Description Model Type kxen RobustRegression Data to be Modeled ksrvmulti2_sqlserver2005 ActveDispositonNoCard Date Cutting Strategy Randomwithouttest DataDescripton None s i s s s Ss S S SS Target Vanable fweditcard bought Weight Variable Optional Nome Compute Decision Tree Enable Au
21. an create it on the fly by clicking the button located on the right of the date field l Note The start date is included in the defined period whereas the end date is excluded from it Define an Aggregate Number of fields that wil be created 1 Hg i din a For Successive Periods Define the number of successive periods you want their length and the starting date using the hyperlinks underlined in blue CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 36 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking l Note The starting date is included in the periods Define an Aggregate Define 2 successive period s of 1 Year starting 1 Year before 2014 02 04 12 00 00 i 13 Click OK 14 Enter the name of the new field in the Name field E Enter the Computed FieldName Name j E Enter a Valid Function Name Cancel 15 Click OK CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 37 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Filtering the Population on a Field Scenario You will now filter the population to keep only the records of interest for your business issue First you need to filter the population to keep only the active dispositions meaning that you will use the aggregate you just created to
22. ate look up table This time range needs to cover the content of your database a SAP InfiniteInsight control panel Create Date Table EE A User Create Date Table Password StartDate 1990 01 01 00 00 01 at wv End Date 2020 01 01 00 00 00 Reuse existing date table 5 Click the button Create Table A message box is displayed indicating that the table is being created 6 Close the Control Panel 3 2 2 Creating the Performance Indicator Scenario A simple performance indicator can be created to visualize the evolution of the credit cards sales each month You will need to create an aggregate counting the number of credit cards bought in the past month and shift it one month forward every month Settings Shifting Aggregate Event Table credit_card Date Column card_date Function count Target Column card_id Aggregation Period 1 Month in the Past Name count_creditcard Period Compute for a period of 730 days which is the number of days between August 1 1996 and August 1 1998 Ending at 1998 08 01 12 00 00 Advanced Filters Create the condition date_day 1 CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 10 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 7 Performance indicator name PI_CountCreditCard vV To Create a Performance Indicator 1 In SAP Infinitelnsight s
23. ce CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 42 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking VI To Copy Objects to Another Metadata Repository 1 Select the objects you want to copy 2 Right click on the selection and choose Copy to another metadata repository in the contextual menu Explorer ta a BE F New Delete Edit View Data Switch View Filtering Options Export KxShell Script gt Export KxShell Script 3j Save as Table or View la Metadata Metadata are stored in the same place as data source Cancel 4 Select aData Type aFolder A Copy to another metadata repository Target Metadata Repository Data Type Text Files Folder C Users natacha yam Documents Other_MDR nie Browse CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 43 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 5 Click Next A summary of the chosen Metadata repository and objects appears 4 Copy to another metadata repository l x Summary Target Metadata Repository Samples Census Text Files All the following objects are going to be transferred Original Name entity_disposition entty_dspositionPopulation a ss5 5 AR_CientGeoDemo i Cancel 6 Click OK to conf
24. ction Support and Integration Documentation of SAP Infinitelnsight documentation CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 48 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 4 2 Describing the Data In order for SAP Infinitelnsight components to interpret and analyze your data the data must be described To put it another way the description file must specify the nature of each variable determining their Storage format number number character string string date and time datetime or date date Type continuous nominal or ordinal For more information about data description see Types of Variables and Storage Formats in the SAP Infinitelnsight User Guide To describe your data you can Either use an existing description file that is taken from your information system or saved from a previous use of SAP Infinitelnsight components Or create a description file using the Analyze option available to you in SAP Infinitelnsight Modeling Assistant In this case it is important that you validate the description file obtained You can save this file for later reuse If you name the description file KxDoc_ lt SourceFileName gt it will be automatically loaded when clicking the Analyze button A Caution The description file obtained using the Analyze option results from the analysis of the first 10
25. d services if any Nothing herein should be 1 P construed as constituting an additional warranty SAP and other SAP products and services mentioned f gine herein as well as their respective logos are trademarks SPETT or registered trademarks of SAP SE or an SAP affiliate r company in Germany and other countries All other product and service names mentioned are the trademarks of their respective companies Please see or additional trademark information and notices
26. e the default date for the prompt VI To Select the Data to be Modeled 1 Inthe Modeler section of Infinitelnsight start panel click the option Create a Classification Regression Model SAP InfiniteInsight Create a Classification Regression Model fh EEE Gisssification Regression Explain and Predict a Target Variable Create a Time Series Analysis Create Association Rules Load a Model Iie a ae Social Recommendation gt Toolkit 2 Check the mode you want to use to select the data to be modeled There are two ways to select the data Usea File or Database Table which allows you to select a standard data source such as database table a text file or a data manipulation created with the SAP Infinitelnsight Data Management feature Use Explorer which allows you to directly select the analytical record and the time stamped population 1 InUsea File or Database Table CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 45 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Inthe list Data Type select the type of the data source Use the Browse button corresponding to the Folder field to select the folder or database containing the data If you want to access a protected database do not forget to enter the username and password when selecting the database Use the Browse button corresponding to the
27. e transfer has been successfully completed 14 Click OK x I D The transfer was successfully completed The transferred entity is now ready to be re used in a new database CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 58 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer 4 2 Use the Transferred Entity in a New Database vV To Use the Transferred Entity in a New Database 1 Inthe SAP Infinitelnsight Start Page click on the option Create or Edit Explorer Objects in the Explorer section ay SAP InfiniteInsight Explorer Create or Edit Explorer Objects Create a Data Manipulation ee Load an Existing Data Manipulation Perform an Event L tion Perform a Sequence Analysis Perform a Text Analysis a z Recommendation Toolkit The panel Connect to Explorer appears 2 Select Data Base in the field Data Type and in the field Database Source the database where the data manipulation has been transferred Connect to Explorer Data Type Data Base Database Source Kervmulti2_sglserver 2005 iL Browse T Metadata Metadata are stored in the same place as data source 1 z CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 59 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer 3 Click Next 4 60 The transferred entity is automatically display
28. eation of new fields application of filters and so on 2 1 6 Temporal Analytical Data Set A temporal analytical data set is a special case of analytical data set It is the product of a time stamped population by an analytical record the result of this operation can be seen as a virtual table containing attributes values associated with identifiers in relation with the time stamp Analytical Record semporal Analytical Data Set cay a ee O a pon i O oO a Be O E CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 5 2014 SAP SE or an SAP affiliate company All rights reserved Essential Concepts In other words a temporal analytical data set contains photos or snapshots of a given list of entities taken ata given time This time can be different for each entity and an entity can be associated with several photos l Note The analytical data sets are used to train predictive descriptive models and to apply these models 2 1 7 Performance Indicator PI The Performance Indicators Pls help organizations achieve organizational goals through the definition and measurement of progress The purpose of defining Pls is to have a common definition of a metric across multiple projects A metric like customer value could easily be defined in several different ways leading to confusing or contradictory results from one analysis to the next Shared Pls ensure consistency acr
29. ed Explorer 20140205 14 42 29 i Metadata Metadata are stored in the same place as data source agd Ges Click Next to edit the entity Perform changes and or click Next to confirm The panel Explorer is displayed listing the newly created entity and the default time stamped population associated with it Explorer CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer www sap com contactsap 2014 SAP SE or an SAP affiliate company All rights reserved No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP SE or an SAP affiliate company The information contained herein may be changed without prior notice Some software products marketed by SAP SE and its distributors contain proprietary software components of other software vendors National product specifications may vary These materials are provided by SAP SE or an SAP affiliate company for informational purposes only without representation or warranty of any kind and b SAP or its affiliated companies shall not be liable for ee errors or omissions with respect to the materials The amp only warranties for SAP or SAP affiliate company products and services are those that are set forth in the express warranty statements accompanying such products an
30. es the maximum value maximum value 11 Inthe list Target Column select the column on which you want to apply the selected aggregation function 12 Onthe line Aggregation Period select the period on which you want to compute the aggregate Shifting Aggregate Aggregation Settings Filters and Pivots Settings Event Table E credit_card Datecoum e o a Aggregate Operation Specification Function Count Target Column card id bel Mumber of fields that wil be created 1 q iT Cancel Previous DP CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 14 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 13 Click the Validate button A pop up window opens allowing you to set a name for the new field you are creating 14 Inthe Name field enter the name you want to give to the field 15 Click the Filters tab Performance Indicator Editor E reids A Merge 7 Fiters Prompts El view Data E SQL Statement Documentation Filters Subsamping Period C Advanced Filter Compute the performance indicator for a period of 60S days ending at 2014 02 04 10 57 11 oe z 16 Select the Period option 17 Use the provided fields to select the period over which you want to compute the performance indicator Performance Indicator Editor C Fields A Merce Y Fiters Prompts El vi
31. ew Data SQL Statement E Documentation C Advanced Filter Compute the performance indicator for a period of 730 days ending at 2008 08 01 12 00 00 gl ra By default SAP Infinitelnsight will compute one point for each line in the table used to create the performance indicator that is either the Calendar table or the time series you have based your indicator on If you want to compute only certain points for example one point every month you will have to create an advanced filter to keep only the dates that interest you As an example if you want a point at the beginning of each month you will create the filter date_day CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 15 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 18 Select the option Advanced Filters 19 Click the button New Condition The expression editor opens 20 Define the filter you want to apply a Define a complex condition using expression editor m l Arithmetic Operators H E Boolean Operators E H Conversion Operators H Date Operators H Miscellaneous Operators Fl String Operators te date_week_label ie date_week_of_month gt je date_week_of_year a due a eld Association Associate by Position r The expression is correct 21 Click the OK button The expression editor closes and the defined condition appears in
32. irm or Previous to change your former selection If you confirm the summary the objects will be saved in the new Metadata Repository 7 Tocheck that the procedure has been correctly carried out select the target folder as the metadata repository in the Explorer panel If the objects have been successfully saved they will be displayed in the objects list Explorer lew DIETS Edit ew Date IE Su fitch View F Jptio z EH port Kushal sopt hl M E e Entity entity_disposition 2012 a 11 17 37 a Time stamped Popul jentity_dispositionPopulat i 2012 10 02 11 17 47 Default Time stamped Population Analytical Record ARL _ChentGeoDema t 2013 01 17 12 09 13 3 4 Creating the First Classification Model You will create a first classification model to see if the analytical record you have created contains sufficient information to predict who will be likely to buy a credit card in the following month CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 44 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 4 1Selecting the Data to be Modeled Settings to Use Select the Data Set Factory Mode Database Source Select the database containing the data Analytical Record select the analytical record AR_ClientGeoDemo ime stamped Population select the time stamped population ActiveDisposition Date Validat
33. isplayed in the list Target Field else the message No Fields Available is displayed Analytical Record Editor E Fields d Merge 7 Fiters 8 Prompts El View Data Zi SQL Statement Documentation Target Field Allowed Type Integer CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 26 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 4 Select the target field in the list Analytical Record Editor Fields Ay Merge 7 Fiters Prompts view Data Gi SQL Statement Documentation Target Table disposition Target Field norinal Remove 3 View Description oo a 5 Click the button New Merge to create the merge Analytical Record Editor E Fields Merge Y Fiters Prompts El view Data E SQL Statement E Documentation CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 2 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 6 Repeat the procedure for each merge you want to create The existing merges are displayed in the upper part of the panel Analytical Record Editor Source Pel Target Table Target Fel Kxld dspositon dsp id entity _depositonPopula ton 3 3 3 Updating the Default
34. ist Events Table Key 8 In the section Aggregate Operation Specification use the Function list to select what kind of operation you want to do in the aggregate Define an Aggregate Aggregation Settings Period Settings Filters and Pivots Settings Table Transaction 5 Date Column date_transaction 7 Reference Table Key Events Table Key id 5 Aggregate Operation Specification The following operations are available CUSTOMER 34 SAP Infinitelnsight Explorer Analytical Data Management v7 0 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 35 Functions Count Sum Average Min Max Exists NotExist S First Last Description computes the number of occurrences computes the sum computes the mean identifies the minimum value identifies the maximum value checks if at least one event exists for the current reference checks if no event exists for the current reference identifies the first occurrence E l Note needs a date column identifies the last occurrence l Note needs a date column Returned Values number of occurrences sum mean minimum value maximum value 0 if no event has been found 1 if at least one event has been found 0 if at least one event has been found 1 if no event has been found value of the first chronological occurrence for the current reference value of
35. ition click the button New Condition Entity Filters Match Al of the Following Match any of the Following The expression editor opens CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 20 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 Youcan either use the offered Functions and Variables by double clicking them or you can enter directly a valid expression The indicator above the Messages field indicates whether the expression is valid or not 4 Define a complex condition using expression editor unior Variables E i Fields H i Field Set H Prompts El Categories ee Extract Categories El Nominal Variables Aga card_type te Functions m l Arithmetic Operators _ Boolean Operators E ie Conversion Operators E ie Date Operators E ie Miscellaneous Operators E ie String Operators H Other Variables jield Association Associate by Position QO Messages The expression is correct a Help OK Cancel 4 Click OK to create the condition The panel listing the conditions is displayed 5 Onceall the conditions you need have been created click OK to validate the filter CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 21 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sel
36. l in Retail Banking 3 3 2 Creating the Analytical Record Scenario Advanced analytics best practice next step consists of creating an overall view of the entities when this entity is a customer it is sometimes called Customer Analytic Record or 360 degree view of the customer or even Customer DNA This view characterizes entities by a large number of attributes the more the better which can be extracted from the database or even computed from events that occurred for each of them The list of all attributes corresponds to what is called an Analytical Record of the entity disposition This analytical record can be decomposed into domains as listed below Demographic Geo demographic Complaints history Contacts history Product history Loan history Purchase history coming from the transaction events Segments Scores Settings to Use Enter AR_ClientGeoDemo as the Analytical Record Name Select entity_disposition inthe list Based on Entity Edit the following attributes in the Merge tab Add information on Target table Source Field Target Field the disposition disposition kxld disp_id the client client client_id client_id geographic and demographic data geodemo home_geocode_id geocode_id the credit card information credit card disp_id disp_id the loan information loan account_id account_id CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0
37. lter has been defined T a filter already exists For the detailed procedure see To Edit a Filter on page 20 7 Click Next to validate the creation or changes on the entity 8 Ifan entity with the same name already exists a dialog box is displayed asking you to choose what should be done Click Yes to save the entity as a new version of the existing one Click No to overwrite the existing entity CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 19 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking The panel Explorer is displayed listing the newly created entity and the default time stamped population associated with it Explorer aq om 9 New Switch View Filtering Options entity_disposition i 2014 02 05 09 59 53 F a Time stamped Popul jentity_dispositionPopulat i 2014 02 05 10 00 01 Default Time stamped Population 2 Performance Indicat PI_CountCreditCard i 2014 02 05 09 55 21 y Metadata Metadata are stored in the same place as data source A Caution Because of technical constraints the default time stamped population must NOT be deleted as long as the entity exists V To Edit a Filter 1 Click the button Edit Filters Edit Entity pb Q Hek Name fentity disposition Descente EEE 7 Edit Fitters The panel allowing you to define filters opens 2 Toaddanew filter cond
38. ndow closes but no action is executed S huge It means that the query will take too much time and resources In that case the behavior of the Continue button depends on the configuration set by the DBMS Administrator for example it can automatically refuse queries that are considered too heavy In any case you should check with them to know the line of action to follow The Explain Mode has not been Configured If your DBMS Administrator has not configured the Explain mode the following pop up opens when you try to access the data k Query Validation E ioj x The Explain feature is not configured Your validation will be required for all queries E No longer request validation for similar queries Show Details um Stop Query YE Continue You need to contact your Administrator who will tell you which action to take and configure the Explain mode If the Administrator validates the execution of the query you may want all queries with the same duration to be executed without validation In that case check the box Do not request validation anymore for similar requests The validation message will then only appear for larger queries This configuration will only be used for the current session when closing SAP Infinitelnsight it will be lost For a permanent configuration see your DBMS Administrator who will find the necessary information in the support document Explain Mode available in se
39. nitelnsight Explorer Analytical Data Management v7 0 30 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 5 Inthe field Prompt Heading enter the text you want to see when the prompt is displayed 6 Click OK to validate the changes A dialog box is displayed asking you if you want to overwrite the existing prompt Conflict with TimeStampPrompt X 7 Click Yes The changes are applied to the prompt Setting Filters on the Time stamped Population Creating a Merge Scenario To get the various identifiers linked to the dispositions account client and the credit card information you need to merge the entity used as basis for the time stamped population with the table containing the dispositions and the table containing the credit card data Settings to Use The table containing the dispositions should already be merged with the time stamped population For the merge with the credit cards table Target table credit_card Source Field KxId Target Field disp_id To know how to create a merge refer to the procedure To Add Fields from Other Tables on page 25 CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 3l 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Creating an Aggregate Scenario To select active dispositions you need fir
40. odedeadassdgeatnodedbeddesdddanebaededdesddctoueededs 54 3 1 Presentation In this scenario you are a data mining analyst in a large retail bank Your role is to translate any business question in this case cross sell into successive data mining functions models You are part of a center of excellence in analytics This group provides support for teams such as the marketing team when developing advanced analytics projects Your project is to increase the number of credit cards subscribed by present customers At the start of year 2008 January 7th the marketing management team decided on an objective to increase the number of credit cards subscribed by present customers of the bank They decided to achieve this objective through the use of advanced analytics Their goal is to engage the marketing operations of the retail bank on the path to competing on analytics 3 2 Creating the Performance Indicator As described in the section Definitions on page 4 gt Performance Indicator PI on page 6 creating a performance indicator will allow you to measure the performance of your company regarding the goal you want to achieve CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 8 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 2 1Creating the Calendar Table Scenario For a given date the Calendar table allows you to get specific inf
41. ormation relative to the date that is complex to compute such as the day of the week the week number in the year the quarter and so on This information is needed to compute the performance indicators Settings ODBC Sources the database in which you have imported the PKDD data 7 Start Date 1990 01 01 00 00 00 z End Date 2020 01 01 00 00 00 VI To Create the Calendar Table 1 Inthe Windows Start menu select the option Programs gt SAP Business Intelligence gt SAP SAP Infinitelnsight gt Control Panel a SAP InfiniteInsight control panel Create Date Table External executables ODBC Sources Excel Files ea D User Create Date Table Password Start Date 2014 02 04 09 19 54 i i End Date 4 02 04 09 19 54 Reuse existing date table ag Creates table 2 Inthe list ODBC Sources select the database in which you want to create the date look up table 3 If the database is password protected enter the user name and password respectively in the fields User and Password ie SAP InfiniteInsight control panel ODBC Sources User Password Start Date End Date 2014 02 04 09 19 54 e CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 9 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 4 Use the fields Start Date and End Date to define the time range that should be covered by the d
42. oss analysts and projects over time The key indicators are agreed upon by an organization and are indicators which can be measured and will reflect success factors The Pls selected must reflect the organization s goals they must be the key to its success and they must be measurable CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 6 2014 SAP SE or an SAP affiliate company All rights reserved Essential Concepts 2 2 Methodology The methodology used in SAP Infinitelnsight for Infinitelnsight Explorer Semantic Layer has been designed around systematization and automation of the creation of temporal analytical data sets The key is to define all attributes describing a customer through expressions that are not based on basic information only but also on a time stamp In other words all attributes are expressions that can be computed in order to get the value of that attribute for a given customer at a given time This methodology is made of four steps Define Entity Define Time stamped Population Define Analytical Record Create Temporal Analytical Data Set 1 First you need to define the entity that will be the main object of the analytical tasks Second you need to define the analytical record that describes the entity s attributes on which predictive and descriptive models can be trained N l Note The first two steps only need to be defined once by a power user
43. pulation select the transfer option Transfer Definition Transfer a Data Manipulation e Would you like to transfer the data or the definition of the data manipulation The panel Data Manipulation Transfer appears 6 Inthe area Target Settings select the Target Database that is the database where you want to transfer your data manipulation inthe field Transfer as enter the name the data manipulation should have in the target data base Data Manipulation Transfer Target Sethngs Target Settings Target Database Cl Browse Transfer as fentity_ disposition 1l Browse Target Metadata Repository Metadata are stored in the same place as data source Table Mapping Crease et sen CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 56 2014 SAP SE or an SAP affiliate company All rights reserved Performing an Entity Transfer 7 Todefine the data manipulation source in the area Table Mapping click Guess If the new database has a similar table with a similar name the application automatically shows the correct Target Table Data Manipulation Transfer Target Settings A Caution The new database must have the similar source table s as the existing database that Is if the Data Manipulation is derived from sources TableA and TableB both TableA and TableB must be migrated to the new database before performing the Data Manipulation transfer 8 Edit the correct field mapping
44. riable stated as a natural order is not actually ordered an error message will be displayed before model checking or model generation Missing the string used in the data description file to represent missing values e g 999 or Empty without the quotes Group the name of the group to which the variable belongs Description an additional description label for the variable Settings to Use Use the Analyze feature Setall the identifiers to nominal CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 50 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 4 3 Selecting the Variables Scenario By default and with the exception of key variables such as Kx ndex all variables contained in your data set are taken into consideration for generation of the model You may exclude some of these variables For the first analysis of your data set we recommend that you retain all variables It is particularly important to retain even the variables that seem to have no impact on the target variable If indeed these variables have no impact on the target variable the model will confirm this In the contrary case the model will allow you to recognize previously unidentified correlations between these variables and the target variable By excluding variables from the analysis based on simple intuition you take the risk of depriving yourself of
45. select only those having at least one transaction in the last 6 months Then you will need to keep only the records for which there is no credit card on January 1 1998 meaning the records for which the card issued date is either empty or is later than January 1 1998 Settings to Use To Keep Only Create the Condition active dispositions TransactionsNumberLast6M gt 0 dispositions without credit cards isNull issued_date issued_date gt KxTimeStamp Vv To Create a Condition 1 Click the button New Condition Time stamped Population Editor Fi Fiels gy Merge Y Fiters Prompts Hl View Data i SQL Statement 2 Documentation Keep Only Records that Match Al of the Following Match any of the Following Filter Condition List The expression editor opens A Define a complex condition using expression editor Variables Fl 2 Fields El Field Set E l Prompts H Categories unctions m l Arithmetic Operators H i Boolean Operators E ie Conversion Operators H Date Operators H i Miscellaneous Operators H String Operators ield Association Associate by Position ka QO Messages SEE CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 38 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 2 Youcan either use the offered Functions and Variables by double clicking them or
46. st to create an aggregate counting the number of transactions there has been for each disposition on the 6 months prior to the date of interest Settings to Use Events Table transaction Events Date Column date_transaction Reference Table Key Events Table account_id Key Function Count Target Column trans_id Period Settings Single Period Period Settings Start Date create a new field KxTimeStampMinus 6M by using the Date Operator Shift Date By Month on the field KxTimeStamp and shift back 6 months You should get the following formula dateAddMonth KxTimeStamp 6 Period Settings End Date KxTimeStamp Name TransactionsNumberLast 6M vV To Create a Field from an Aggregate 1 Select the Fields tab Time stamped Population Editor elds IA Merge y Fitters 8 Prompts view Data E SOL Statement Documentation t F GO vD New Info Options KxTimeStamp KxTimeStamp nominal DateTime Function z g kid Kxid homnal Integer lentity_disposi 5 icard_id card_id jcon Gnueus Integer kxenodbcoredi 3 k d_type lcard_type nominal String lovenodbcoredi a jcard_date card _date leonttinuous DateTime ocenodbcoredi a M Display only visible fields CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 32 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 2 Inthe menu select the option New Aggregate
47. t a variable in the screen section arget s Variable s and click the button lt to move the variables back to the screen section Explanatory Variables Selected vV To Select a Weight Variable On the screen Selecting Variables in the section Explanatory Variables Selected left hand side select the variables you want to use as a Weight Variable oN l Note On the screen Selecting Variables variables are presented in the same order as that in which they appear in the table of data To sort them alphabetically select the option Alphabetic sort presented beneath each of the variables list CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 51 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 4 Click the button gt located on the left of the screen section Weight Variable middle right hand side Weight Variable Ss The variable moves to the screen section Weight Variable Also select a variable in the screen section Weight Variable and click the button lt to move the variables back to the screen section Explanatory Variables Selected vV To Exclude Variables 1 On the screen Selecting Variables in the section Explanatory variables selected left hand side select the variable to be excluded l Note On the screen Selecting Variables variables are presented in the same order as that in which they appear in the t
48. tart panel select the option Create or Edit Explorer Objects in the Explorer section 2 Click the Browse button to select the database in which your data are stored Connect to Explorer Data Type Data Base Y Database Source l Browse Metadata Metadata are stored in the same place as data source oe Cancel Previous 3 If the database is protected enter the username in the User field and the password in the Password field fe Data Selection Select a Database B Excel Files wen i e B MS Access Database B Visio Database Samples B dBASE Files 4 Click OK to validate your selection 5 Click Next to display the panel Explorer Explorer ap A aJ E a Switch View Filtering Options i Metadata Metadata are stored in the same place as data source CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 11 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 6 Inthe menu select the option Performance Indicator gt Based on the Calendar Table Explorer QB Entity amp Analytical Record do Time stamped Population 4 Temporal Analytical Data Set E Cross Table With Dates Metadata Metadata are stored in the same place as data source owe Cancel Previous The Performance Indicator Editor opens Performance Indicator Editor R Display only visible fields 7
49. the Filter Condition List Performance Indicator Editor E Feks A Merge T Fiters 8 Prompts E View Data i SQL Statement E Documentation Y Fiters Subsamping Period Advanced Filter Keep Only Records that Match Al of the Following Match any of the Following Filter Condition List T Ag Cancel Previous CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 16 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 22 Click the tab View Data to display the performance indicator as a graph Performance Indicator Editor E Fields Merge Y Fiters G Prompts El View Data ig SQL Statement Documentation Data Statistics BS Graph ld SA eA Sep 108 Nowl08 Jan 1907 Mar 1907 May 1807 JukiBe Sep 1907 Now 1807 Jan 1S06 Maig May 1888 Jub igea keen _ date a count_creditcard CNT_card_id 23 Click Next the panel Save and Export is displayed Save and Export W Save Data Manipulation Mame i Description E Save as Table or View Save as a Tabie Save asa View Name of the Table View l KxShell Script Export KxShell Script Saving Location 24 Inthe field Data Manipulation Name enter a name for your performance indicator 25 Use the Description field to enter a description of the performance indicator 26 Click Save CUSTOMER SAP Infinitelnsight
50. the expression KxTimeStamp lt issued_date amp amp issued_date lt dateAddMonth KxTimeStamp 1 Computed Field Name enter creditcard_bought VI To Create the Variable to be Used as the Target 1 Inthe Explorer panel select the time stamped population to which you want to add a target 2 Click the U sutton Edit The Filtered Time stamped Population panel is displayed 3 Click the button Edit Filters The Time stamped Population Editor opens In the menu select the option Expression Editor Time stamped Population Editor C Fields A Merge Y Fitters 8 Prompts view Data E SQL Statement E Documentation jy 9 ev Info Options J e ee e e e a M Display only visible fields oo cs The Expression Editor opens CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 AO 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 41 5 You can either enter a valid expression directly in the upper text field or you can insert functions variables field sets and prompts by double clicking to insert them at the cursor location Expression Editor KxTimeS tamo lt issued date 8A issued _date lt dateAddMonthiKxTimeStamp 1 l Il l s Functions Variables T Arithmetic Operators m J Fields H E Boolean Operators H l Field Set H E Conversion Operators E Prompts 1 Date
51. the last chronological occurrence for the current reference In the list Target Column select which column to use for the aggregate Define an Aggregate Aggregation Settings Period Settings Filters and Pivots Settings Table Transaction Lel Date Column Idate_transaction Reference Table Key Aggregate Operation Specification Function Average ae Mumber of fields that wil be created 1 CUSTOMER 5 Events Table Key account id Lel Not compatible with Average Not compatible with Average Z Previous SAP Infinitelnsight Explorer Analytical Data Management v7 0 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 10 If you want to define a time window for your aggregate click the tab Period Settings Define an Aggregate Aggregation Settings Period Settings Filters and Pivots Settings Singe Pernod C Start Date Field Constant Argument With Date F we femmeseme A 1 f Field Constant Argument With Date mee mest OOOO H 12 Select if you want to define a Single Period or several Successive Periods Fora Single Period Select a Start Date and an End Date for the period over which you want to aggregate the data You can either use an existing field Field a constant Constant or a prompt Prompt If the field you want to use does not exist you c
52. to selection w aa GE The name of the model is automatically filled It corresponds to the name of the target variable followed by the underscore sign _ and the name of the data source minus Its file extension 1 Note The screen Summary of Modeling Parameters contains an Advanced button By clicking this button you access the screen Specific Parameters of the Model where you can set the polynomial order of the model to be generated For more information see the K2R section of the SAP Infinitelnsight User Guide CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 53 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 5 Analyzing the Model Results If you look at your model results you can see that the performance indicators are not very high The quality of your model represented by the KI can be improved by adding information to the analytical record The analytical record you have created contained only geographic and demographic data however you have a lot of information available on the customers behavior such as the amount withdrawn from an account or the amount checked in the account the balance at the end of each month and the tenure of the account You could also add the product history with the number of loans taken in the last year for example and the loan history However this first model already reveals some
53. ts reserved Welcome to this Guide 1 Welcome to this Guide 1 1 What this Document Covers This document presents the new data manipulation functions offered by SAP Infinitelnsight through a scenario The first section provides definitions of essential concepts as well as the methodology to be applied when using these new functions The second section guides you through the steps defined in the methodology by using a practical scenario as an example CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 3 2014 SAP SE or an SAP affiliate company All rights reserved Welcome to this Guide 2 Essential Concepts IN THIS CHAPTER EEE pe ese iececn E EE classic etm EEEN ciel EE ASA erates ecto dence A E xc eara ine ote dio E E EN NACL glols 0 0 eere enters arse neler nee eae Me Merete ere E E E eer eee ey ere meee oe es eee E eer eee 2 1 Definitions 2 1 1 Infinitelnsight Explorer Semantic Layer The Infinitelnsight Explorer Semantic Layer as defined by SAP Infinitelnsight is made of three elements the Data Manipulation features offered by SAP Infinitelnsight such as filters join attributes new attributes computation aggregates performance indicators definition the SAP Infinitelnsight analytical data set methodology see Methodology on page 7 the metadata management which allows you to store share and easily re use the data descriptions 2 1 2 Entity An entit
54. ult the attributes are associated to a domain named after the table they are originated from 7 Click the Next button to validate the creation or edition of the analytical record 8 Ifan analytical record with the same name already exists a dialog box is displayed asking you to choose what should be done Click Yes to save the analytical record as a new version of the existing one Click No to overwrite the existing analytical record CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 24 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking vV To Add Fields from Other Tables Create a Merge 1 Select the Merge tab Analytical Record Editor Target Field No Fields Available KxTimeStamp nominal jentity_dispositi 0 H Target Field No Fields Available CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 25 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking 3 Select the table to be joined in the list Target Table Analytical Record Editor Fields Ay Merge 7 Fiters Prompts i view Data Gi SQL Statement Documentation Target Table Select Target Table No Feld Available w View Description If the selected table contains fields corresponding to the source field type they are d
55. y is the object of interest of any analytical task It can be a customer a product a store and is usually identified with a single identifier that can be used throughout the data repositories Entities are usually associated with a state model describing the life cycle of such an analytical object of interest l Note This is a technical constraint entities MUST be uniquely identified CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 4 2014 SAP SE or an SAP affiliate company All rights reserved Essential Concepts 2 1 3 Time stamped Population A time stamped population is a set of entities at a particular point in time For example all customers as defined by customer_ID the entities as of January 1 2008 the time stamp It is a list of pairs lt identifiers time stamps gt the semantic meaning of such a construct can be associated with snapshots of the entities and a given time in general terms a given entity may be represented at different time stamps in a single time stamped population 2 1 4 Analytical Record An analytical record is a logical view of all attributes corresponding to an entity An analytical record may be decomposed into domains that group attributes related to each other For example in CRM an analytical record can have a demographic domain and a behavioral domain 2 1 5 Analytical Data Set An analytical data set is a data set generated by manipulating data through merges cr
56. you can enter directly a valid expression The indicator above the Messages field indicates whether the expression is valid or not a Define a complex condition using expression editor greater TransactionsNumberLast6M O Variables po pg nemn da TransactionsNumberLast6M iha account_id Ean g aggregate_CNT_account_id g card_date iha card_id ily card_type he client_id i hey dient_type i a disp id H Field Set H Frompts H Categories Functions l 2 Boolean Operators iha And iha Does not Start with cis Equal af M Greater than ty Greater than or Equal to fy Is Null s Is True idg Is in List fg Is in Range fy Is not Null i a Is notin List fy Is not in Range ies Like jield Association Associate by Position QO Messages The expression is correct 3 Click OK CUSTOMER SAP Infinitelnsight Explorer Analytical Data Management v7 0 39 2014 SAP SE or an SAP affiliate company All rights reserved Application Scenario Segmented Cross sell in Retail Banking Creating the Target Scenario You want to create a model able to predict who is likely to buy a credit card in the following month The time stamped population you have created contains only dispositions that have no credit card associated on January 1 1998 so you need to create a target indicating who has bought a credit card between January 1 1998 and February 1 1998 Settings to Use Inthe Expression Editor type

User Guide - SAP Service Marketplace

Contents

Download Pdf Manuals

Related Search

Related Contents