Cognos case study raiffeisen bank - marijana tomaševic

39
RBRS Cognos Reporting IT development department DWH&BI Unit Belgrade, September 2011 Marijana Tomašević

description

 

Transcript of Cognos case study raiffeisen bank - marijana tomaševic

Page 1: Cognos case study    raiffeisen bank - marijana tomaševic

RBRS Cognos Reporting

IT development departmentDWH&BI Unit

Belgrade, September 2011Marijana Tomašević

Page 2: Cognos case study    raiffeisen bank - marijana tomaševic

2

Agenda

� DWH, BI and Cognos in Raiffeisen Bank

� Data Modeling

� Cognos Applications in RBRS

� Specific Cognos Features In Use – Tips and Tricks

� Conclusion

Page 3: Cognos case study    raiffeisen bank - marijana tomaševic

3

DWH, BI and Cognos in Raiffeisen Bank

Source Source databasesdatabases

Data marts & Data marts & Analytical SystemsAnalytical Systems

Other Other sourcessources

FrameworkManager

FrameworkFrameworkManagerManager

DWHDWH

CBSCBS

DWHDWH

Data MartsData Marts

AnalyticalAnalyticalSystemsSystems

ReportingReporting

GDWH in Raiffeisen Banka is a Raiffeisen group wide Oracle solution with local extensionslive since December 2004.

Page 4: Cognos case study    raiffeisen bank - marijana tomaševic

4

For 3 Years in-house team of 2 - 3 developers doing both, DWH (mapping, ETL, operations, testing etc) and BI matters (metadata preparation, standard reporting, data mart preparations, etc).

IT DWH & BI Unit created with two teams: DWH (3) and BI (3) in 2008.

DWH, BI and Cognos in Raiffeisen Bank

Q2Q2

Cognos 8 BICognos 8 BICognos MR3Cognos MR3

20020055 20092009 20112011

Q2Q2 Q4Q4

Cognos 8.4Cognos 8.4

20072007

Page 5: Cognos case study    raiffeisen bank - marijana tomaševic

5

With increasing number of diverse reporting requirements our motivation was to:

� Create data marts and Cubes that would satisfy the majority of requirements, in order to facilitate development and raise data quality with one source for many reports/departments in the Bank.

� Educate Cognos power users to create ad-hoc queries and reports by themselves from prepared data marts.

DWH, BI and Cognos in Raiffeisen Bank

Page 6: Cognos case study    raiffeisen bank - marijana tomaševic

6

BI Initiative For IT & Business Users

Cognos training by our local Cognos partner in December of 2007:� Cognos 8 BI Reporting and Modelling� Cognos 8 BI Analysis and OLAP Modelling

Locally organized workshops introducing Reporting & OLAP and Data Mining tools’ capabilities and added value for daily user operations.Gap analysis and functional specification for Data Marts for different units performed.

Page 7: Cognos case study    raiffeisen bank - marijana tomaševic

7

Currently, there are three active Cognos servers:

� One with version of Cognos 8.2, test and production server at the same time.

� Two with version of Cognos 8.4, one test and one production server.

System Architecture

8.4 8.48.2

Page 8: Cognos case study    raiffeisen bank - marijana tomaševic

8

Framework Manager used by IT for preparing reporting metadata.

Query Studio is not used in RBRS as it turned out not to be acceptable choice with huge amounts of data they deal with.

Report Studio is used by Report Authors when creating Ad-Hoc and classic reports.

PowerPlay is used by certain departments for exploring data received from HO.

Analysis Studio was explored a little from our side, but is currently not used by business users.

Cognos Tools In Use

Page 9: Cognos case study    raiffeisen bank - marijana tomaševic

9

RBRS Bussiness Departments – Cognos Users

Users of Cognos portal and Cognos Applications are members of different departments within RB :

� Back Office Departments (Accounting...)

� Analysis Departments (Business Analysis...)

� Sales Departments (Retail Sales Finance...)

� Reporting Departments (Controlling...)

� Risk Departments (Market Risk...)

Page 10: Cognos case study    raiffeisen bank - marijana tomaševic

10

Satisfaction with Cognos tools can be seen through continually increasing number of active users.

Number of Users

N um b er O f D istinct Cognos Users

0

10

20

30

40

50

60

70

20

08

.06

20

08

.08

20

08

.10

20

08

.12

20

09

.02

20

09

.04

20

09

.06

20

09

.08

20

09

.10

20

09

.12

20

10

.02

20

10

.04

20

10

.06

20

10

.08

20

10

.10

20

10

.12

20

11

.02

20

11

.04

20

11

.06

20

11

.08

Page 11: Cognos case study    raiffeisen bank - marijana tomaševic

11

Agenda

� DWH, BI and Cognos in Raiffeisen Bank

� Data Modeling

� Cognos Applications in RBRS

� Specific Cognos Features In Use – Tips and Tricks

� Conclusion

Page 12: Cognos case study    raiffeisen bank - marijana tomaševic

12

Data modeling

First Cognos FM models were based on standard DWH model.

Experience with such models indicated that with quite huge amount of data that RBRS has, choosen approach was not good enough. Cognos had problems with calculating data required for reports.

Present practice is to create Data Mart for FM model.

So, we switched to generating Data Marts precalculated and suited to user needs. In that way, Cognos Applications are used to show prepared data rather than to calculate data.

Page 13: Cognos case study    raiffeisen bank - marijana tomaševic

13

Data Marts – Develop Once Use Many Times

� Created to fulfill requirements of different business segments.

� Created on monthly or daily basis.

DWHDWH

Daily Data MartsDaily Data MartsMonthly Data MartsMonthly Data Marts

Page 14: Cognos case study    raiffeisen bank - marijana tomaševic

14

Agenda

� DWH, BI and Cognos in Raiffeisen Bank

� Data Modeling

� Cognos Applications in RBRS

� Specific Cognos Features In Use – Tips and Tricks

� Conclusion

Page 15: Cognos case study    raiffeisen bank - marijana tomaševic

15

Cognos Applications

� Cognos as front-end for MIS System

� 5 major Cognos applications: KPI portal, RORWA Tool, Deposit DM, DQI Engine and Cost Accounting Application

� 7 Data Marts

� 5 Cubes

� > 200 locally developed reports, 80 centrally provided by HO.

Page 16: Cognos case study    raiffeisen bank - marijana tomaševic

16

Cognos Applications – RORWA Tool

Page 17: Cognos case study    raiffeisen bank - marijana tomaševic

17

Date, analytical and instance selection

Top 4 employees by difference from

benchmark value

Top 4 empoyees by improvement from last month value

KPI Portal Page - Details

Page 18: Cognos case study    raiffeisen bank - marijana tomaševic

18

Historical KPI

KPI Portal Page – History Overview

Page 19: Cognos case study    raiffeisen bank - marijana tomaševic

19

� DWH, BI and Cognos in Raiffeisen Bank

� Data Modeling

� Cognos Applications in RBRS

� Specific Cognos Features In Use – Tips and Tricks

� Conclusion

Agenda

Page 20: Cognos case study    raiffeisen bank - marijana tomaševic

20

FM defined data security concept. Data is filtered by membership of Cognos users groups within defined Cognos roles created based on customer segmentation.

Implementation Of Data Protection

Page 21: Cognos case study    raiffeisen bank - marijana tomaševic

21

When using query items from secured query subject in creating a report, where clause contains security check that returns only data allowed for user running or creating a report to see.

Implementation Of Data Protection

Page 22: Cognos case study    raiffeisen bank - marijana tomaševic

22

Example Of Visibility Of Report Navigator

Simmilar security approach implemented in regular reports is used to show different links to different reports, so depending on user logged, Deposit Navigator interactively changes.

Page 23: Cognos case study    raiffeisen bank - marijana tomaševic

23

Example Of Report Pages Rendering

The same report renders differently depending on user logged, and parameters supplied.

Page 24: Cognos case study    raiffeisen bank - marijana tomaševic

24

Cognos Features In Use – Drilling

Detail records are exported to excel files for further analysis

Life before Cognos: complex gigantic excels or access databases with all data for analysis/troubleshooting. With Cognos: targeted reports with drill through capabilities for narrow search results and faster analysis.

Page 25: Cognos case study    raiffeisen bank - marijana tomaševic

25

Native cube vs user expected drilling

Cube report shows summaries accross different dimensions. Cognos enables drilling up and down within a cube.

For user convenience and targeted analysis, detail records are exported to excel files.

Page 26: Cognos case study    raiffeisen bank - marijana tomaševic

26

When distributing report results via email, few scenarios exist:

� Internal DQ reports scheduled to be run after daily DWH load. Excel files sent via email.

� Bursted daily reports with details of products for customers are mailed to relationship mangers in charge. Their superiors receive same reports with the data of all RELMs they are responsible for.

� Some users receive only notifications that reports are run, results are saved on Cognos server, and can be downloaded via link indicated in email.

Distribution Of Reports

Page 27: Cognos case study    raiffeisen bank - marijana tomaševic

27

Scheduling of a single report

� Start by using triggers� Check every few hours whether certain conditions are met and

then start (conditions ranging from finishing the creation of the data mart to finishing the import of external files recieved through file transfer software).

Run a set of reports (business request)

� Jobs and events are used to runset of reports every working day bytrigger after certain conditions are met.

Cognos Features In Use – Scheduling Of Reports

Page 28: Cognos case study    raiffeisen bank - marijana tomaševic

28

Example of a daily reconciliation report.

Scheduling Of a Single Report

Page 29: Cognos case study    raiffeisen bank - marijana tomaševic

29

Trigger Implementation

DWHDWH

Data Stage job calls NRPE script located on Data Stage Server.Script contains address of Cognos Server and name of the Cognos trigger.

Page 30: Cognos case study    raiffeisen bank - marijana tomaševic

30

Trigger Implementation

When scheduling a report, user should only specify the name of the trigger.

Trigger is located in a predefined folder on the Cognos server.

Page 31: Cognos case study    raiffeisen bank - marijana tomaševic

31

Cognos Features In Use – Bursting of Reports

Page 32: Cognos case study    raiffeisen bank - marijana tomaševic

32

� DWH, BI and Cognos in Raiffeisen Bank

� Data Modeling

� Cognos Applications in RBRS

� Specific Cognos Features In Use – Tips and Tricks

� Conclusion

Agenda

Page 33: Cognos case study    raiffeisen bank - marijana tomaševic

33

Cognos enables us to use one well defined single-point-of-truth data source to:

� Secure and protect our data

� Analyze data with ease

� Create personalized reports without IT resources

� Receive results in most convenient and timely manner

Benefits

TRERWA

Page 34: Cognos case study    raiffeisen bank - marijana tomaševic

34

Questions

Page 35: Cognos case study    raiffeisen bank - marijana tomaševic

35

Appendix

Page 36: Cognos case study    raiffeisen bank - marijana tomaševic

36

Modeling Approach

Best practice relational model is made up of few layers, for instance:

� A “Database View ”, where the database objects (tables, views) and their relationships are mapped in an almost one-to-one direct manner to the ReportNet model.� A “Business View ”, where query elements are structured and named in a more business oriented manner, hiding the database structures (tables, views) and their relationships from the user.

Database View is used as the basis for the Business View.

Relationships are always modeled in FM when models are used by power users.

Page 37: Cognos case study    raiffeisen bank - marijana tomaševic

37

Cognos Authors who use DWH based FM model sometimes (from their own words, 1-2 times a month) expirience prolonging of run time of Ad-Hoc reports, which is usually resolved by analysing Cognos generated SQL, and reconstructing the report with our help.

This is probably due to refreshed statistics on some of the DWH tables, but we have never investigated the reasons for such a behaviour.

Performance Issues Cognos 8.2

Page 38: Cognos case study    raiffeisen bank - marijana tomaševic

38

When in process of creating Cognos Cube, Cognos can run out of temporary space.

While calculating data for cube, temporary working file can require up to 100GB, although cube ends up with size of less than 2GB. Thisdepends on number of categories and on number of values within categories. Error thrown hascode TR0114*. IBM advises: To estimatehow much disk space you need to createa cube, multiply the size of all data sourcesby 3.5. To locate the temporary work files,check the Directories tab of the Preferencesproperty sheet.

Performance Issues Cognos 8.2

*Transformer cannot write in the model temporary file.Please check if there is enough free disk space in the temporary directory.

Page 39: Cognos case study    raiffeisen bank - marijana tomaševic

39

Problems with local proccessing that occures on Cognos Server.Cognos is creating temporary file in folder which can be located on the path C:\Program Files\cognos\c8\temp.

When testing report which was combining data from two different databases, temporary file created has reached size of 10GB, and the error UDA-TBL-0004* was thrown.

Resolved by optimizing queries to deal with onlynecessary rows and columns, thus reducingnumber of rows processed.

Performance Issues Cognos 8.4 Test Server

*There was a Write error while processing a temporary file.