Thursday, May 30, 2013

Class diagram, Sequence, Use case, Activity, Component, and system diagram of warehouse based banking analytics

Outline
banking analytics is a warehouse based project. This is a project of analysis of transactional data in banks and tried to analyse the three objectives

  1. fraud detection
  2. churn analysis 
  3. customer relationship management
This project entitled “Warehouse based intelligent banking transaction analysis system” aims to develop a data warehouse with intelligent enough to analysis the day to day transaction occurring in the banks.  The key aims of the project  is to make the banking transaction more reliable and secure which has been a major problems in today’s age of increasing information technology. Furthermore this project aims to withstand a competitive market and bring itself in a strong position with the analysis of customer behavior and activities.


  •  1      System Architecture







The application system architecture consists of four blocks including the data sources, data warehouse servers, OLAP servers and reporting & data mining block. The functionality of each of the blocks is illustrated below:

                        Block –I: Operational & External Data Sources


For the implementation of a data warehouse & business intelligence system, the availability of reliable and actual data sources is essential and the most important without which the information reported, mined and forecasted may not be fruitful. For our system, the vendor for the operational & the external data sources is Nepal Investment Bank. The bank provided the bank’s customers’ profile and transaction databases in various format such as .txt and .sql. These data sources are flat files and need to be converted in multi-dimensional format for OLAP operations.

                         Block – II: Data Warehouse Servers

This block contains the staging area, warehouse database servers & metadata repository. There is physical data movement from source database to data warehouse database. Staging area is primarily designed to serve as intermediate resting place for data before it is processed and integrated into the target data warehouse. This staging are serves many purposes above and beyond the primary function:
                    The data is most consistent with the source. It is devoid of any transformation or has only minor      
                     format changes.

                     The staging area in a relation database can be read/ scanned/ queried using SQL without the  
                     need of logging into the source system or reading files (text/xml/binary).

                       It is a prime location for validating data quality from source or auditing and tracking down data  
                      issues.

                    Staging area acts as a repository for historical data if not truncated.

The next component is a warehouse database server that is almost always a relational database system. Back-end tools and utilities are used to feed data into the bottom tier from operational databases or other external sources (such as customer profile information provided by external consultants). These tools and utilities perform data extraction, cleaning, and transformation (e.g., to merge similar data from different sources into a unified format), as well as load and refresh functions to update the data warehouse.
This block also contains a metadata repository, which stores information about the data warehouse and its contents.

                     Block – III: OLAP Server

The middle block is an OLAP server that is typically implemented using either
(1) a relational OLAP (ROLAP) model, that is, an extended relational DBMS that maps operations on multidimensional data to standard relational operations; or
(2) a multidimensional OLAP (MOLAP) model, that is, a special-purpose server that directly implements multidimensional data and operations.

                Block – IV: Reporting & Data Mining Tools

The front-end client layer in data warehousing is the presentation phase which contains query and reporting tools, analysis tools and data mining tools for trend analysis, fraud detection and customer churn behavior analysis. The reporting tool that we’ve used for this purpose is Oracle Business Intelligence Enterprise Edition (OBIEE) 11g.

For providing the analytical result, we will be using some of the Online Analytical Processing (OLAP) operations such as slicing & dicing, roll up & roll down and pivoting. The analytical results will be provided in a multi-dimensional view using OLAP Cube Technology projected to assist decision makers such as visualization with comparison to different dimensions e.g. locations, time etc.

For trend & prediction analysis featuring churn analysis and CRM, we will be using some of the data mining algorithms such as CART, C5.0 & Rule based algorithms. Transactions made by fraudsters using counterfeit cards and making cardholder-not-present purchases will be detected through methods which seek changes in transaction patterns, as well as checking for particular patterns which are known to be indicative of counterfeiting. 


  • 2     USE CASE DIAGRAM

                          loading datawarehouse

use case diagram of loading datawarehouse


As shown in diagram above, the loadDataWarehouse subsystem contains two actors one primary actor- Database Administrator and the other supporting actor- Bash shell ETL module. Each of the actors show their one behaviour in context to the system. The relationships between the actors and the use case scenarios have been shown by the association lines and those between use cases have been shown by the dependency lines with appropriate stereotypes shown within guillemets. The specification of the actors with main success scenarios and other alternate scenarios have been shown as follows:

                                         Specification of Actors

element
Details
Description
Database Administrator is the primary actor who is responsible for the extraction, transformation and loading of the data to the staging area. This actor is also responsible for the handling of new transaction data.


element
Details
Description
Bash Shell ETL Module is the supporting actor who is responsible for actual extraction, transformation and loading of the data to the staging area. This actor is also responsible for the loading of the target area and handling updates, i.e, refreshment of the warehouse.






As shown in the figure above, the use case diagram modeling the report generation and data mining subsystem consists of two actors- the database administrator and the end users which are the executives belonging to the tactical level of the management in the bank. The relationships between the actors and the use case scenarios have been shown by the association lines and those between use cases have been shown by the dependency lines with appropriate stereotypes shown within guillemets. The specification of the actors with main success scenarios and other alternate scenarios have been shown as follows:

 Specification of Actors


element
Details
Description
Database Administrator is the supporting actor who is responsible for the handling of different multidimensional processing of data. This actor is also responsible for providing the supporting environment for the end users.


element
Details
Description
Bank Manager is the primary actor who is responsible for the generation of different analytical reports, dashboards and also perform the fraud detection and churn prediction including the CRM analysis provided by the subsystem.








  • SEQUENCE DIAGRAM
                    system sequence diagram




Data source are heterogeneous may have different storage file. Data is extracted from the distributed heterogeneous data sources and stored in DSA (Data Staging Area) and is used in the next level of ETL process, i.e., cleaning and transformation. The cleaned and transformed data is then loaded into the data warehouse

                          report generation sequence diagram








As in the figure there is three class OBI, OBIBackend, DataWarehouse and BankManger is user.  Synchronous message passing is done between classes which is shown by the solid and dotted line. OBI get the request from user and OBI send request to OBIBackend to generated report. OBIBackend gets necessary data from dataWarehouse class and generated report is send to OBI and report is displayed to the user.


ACTIVITY DIAGRAM


activity diagram of reporting system



The facts and dimension from the warehouse are arranged in snowflake schema. OLAP cube are formed through which operations like slicing, dicing, roll up & roll down are performed and then report are generated according to those fact tables.


CLASS DIAGRAM






COMPONENT DIAGRAM

The above component diagram displays the major components involved in this sub system and communication between them. The ATM Machine component denotes the ATM Machine where a transaction takes place, the users must authorize themselves before performing any transaction. The ATM transaction is passed to the Bank Database component via LAN interface. The data stored in Bank Database are accessed by ETL tools component with necessary credentials and authority to perform extraction, transformation and loading of the raw data from Bank Database.  Finally the transformed and refined data are loaded into the data warehouse.





Above component diagram displays the components involved in report generation using OBIEE tool. The data from Data Warehouse component are fed into the Oracle BI server. Oracle Bi server provides efficient processing of data and structure information intelligently. It uses metadata to direct processing. The Metadata Repository component stores the metadata used by Oracle BI server. The query client component contains two components, Analysis Editor Component and Dashboards component. Analysis Editor Component contains set of graphical tools that enable users to build, view, and modify analyses that provide analytical information. Dashboards display results the analyses and other items. The Administrator tool component exposes the Oracle BI repository as three separate panes of layers: Physical, Business Model and Mapping and Presentation






3 comments:

  1. Amazing stuff here, very nice to read this kind of blogs, thanks for sharing it. . .
    Packers and Movers delhi.

    ReplyDelete
  2. Thanks this info was really helpful! I used a website called Lucidchart to create my own component diagram and it was really easy to understand. If you use diagrams often you should check it out!

    ReplyDelete
  3. Thanks for posting unique article.
    For more related information visit us :-
    Warehouse management solutions.

    ReplyDelete