A Solution for Business Analytics

2008-06-04T21:28:00.000-07:00

Table of Contents
1. Overview
2. The Architecture
3. The Data Sources
4. The ETL
5. The Reports
6. Monitoring
7. Conclusion

1.0 Overview
The Business Analytics solution implementation is a Web-based decision support solution that collates data from various sources that exist within an organisation (i.e. it builds a data warehouse) and performs the analysis required on the data by converting it into meaningful information. The information is then presented in a customised format to the end users.

With this solution users can:
• Gather information from various sources such as existing software and databases and transform this data into a management dashboard that will alert key managers of the health of the business.
• View data in the form of graphs or multi-dimensional reports. Easy to use drill-down features allow for zooming into any aspect of the business instantly.
• Review targets vs. achievements of business performance so as to analyse why certain business areas perform well, while others do not.
• Download graph data and import it directly into Microsoft® Excel®. Graphs can also be pasted into Microsoft® Word® or Powerpoint®.
• Personalise their dashboards and reports.
• The front-end is a Web browser with an intuitive interface allowing users to become productive in a short span of time.
• Administrators can set up the system and create user roles, Key Performance Indicators (KPIs) and multi-dimensional views of KPIs. A key role of the administrator is to ensure that data access is limited to the information needs of a particular user role.

Fig 1. High Level Representation of the Business Analytics Solution

2.0 The Architecture
The solution adopts a multi-layered architecture to deliver a highly secure and performance-oriented e nvironment for providing information. The three layers are the Data Access Layer, the OLAP technology layer and the Presentation Layer.

2.1 The Data Access Layer
The Data Access layer is used so that the organization can access and extract data from different IT applications. Every company uses several software applications and each software application has its own data sources. Consolidating these sources to obtain comprehensive data is often a concern when Key Performance Indicators (KPIs) need to be tracked.
The solution has automated the software processes that load KPIs and their corresponding data into its own warehouse. The warehouse stores different KPIs as well as user information. User information includes user views, dimensions and facts, saved dashboards, and commonly used queries. This information is stored in the Configuring Tables.
Every time a user starts to interact with the system an authentication process is run to ensure the user is valid. Users are connected to the engine only after the authentication. If a pre-catalogued query is made, the system displays the results of the query. If a new query is to be created, the system allows the user to configure the query in the fashion required.

2.2 The OLAP Technology Layer
The second layer or the high performance analytical engine layer generates graphs, reports, tracks KPI, and safeguards information. Most commonly accessed information is pre-catalogued and this can be easily presented with high speed.
OLAP services enable a user to create multi-dimensional views. The Configurator helps to configure filters, preferences for dashboard and accessibility of users.

2.3 Presentation Layer
The Presentation layer or the User Interface layer supports several options including access through a simple web browser as well as pervasive devices like PDAs to ensure anytime, anywhere access to business information.
The User Interface layer and the Technology layer are separated by a secured firewall. Only users with access rights can access the technology layer.

3.0 The Data Sources
The data for the Business Analytics system is provided from two sources, one being the log files of applications and the other through a web service or an XML feed. The data is provided on a daily basis after the entire day’s data is available. The data feed is scheduled as a batch process.

3.1 Log Files
Log files of applications contain application data which forms the input to the reporting system. These applications do not have any data in databases that needs to be part of the reports, hence the data is extracted from the log files. The required log files are collated and provided to the system at pre-defined locations from where the Business Analytics ETL process loads the data into its own database.

3.2 XML Feeds / Web Services
Most applications have data in their databases that requires to be made available for reporting. This data is provided to the Business Analytics system in the form of an XML feed. The web services create the XML feed of the entire days data for particular applications and make it available to the ETL process of the Business Analytics System.

Fig 2. The Data feeds for the ETL Process

4.0 The ETL Process
4.1 The data input
The ETL process receives the data which is provided either in the log files or through the XML feeds from the various applications.

4.2 The staging db
The data is read from the XML files / Log files and verified for consistency and completeness. This is achieved using the protocol for the data exchange. The data is then extracted and loaded into the staging area where it is processed for loading into the warehouse.

4.3 The warehouse
Once the data is loaded into the staging area and transformed, it is then loaded into the warehouse where it is made available for reporting.

Fig 3. The ETL Process

5.0 The Reporting
The dimensions and the measures for the reports are configured in the system depending upon the business needs of online service for which the business analytics is required. This could differ from application to application and the required reports, dimensions and measures are captured during the business requirements and specifications stage and configured in the system according to these specifications.

6.0 The Monitoring
The system has alerts built in that are designed to help in the monitoring process. The alerts cover the data input from the log files and the XML feeds and the ETL process itself. The alerts are configured as notification emails and sms to the monitoring team.

Fig. 4. The alert mechanisms for monitoring the system

7.0 Conclusion
The document described a Business Analytics solution implementation which is a Web-based decision support solution that collates data from various sources that exist within an organisation (i.e. it builds a data warehouse). It performs the analysis required on the data by converting it into meaningful information. The information is then presented in a customised format to the end users.