Dashboards for Learning management system

In Resume
Project	Dashboards for Learning management system
Customer	Chrysler LLC & FIAT
Period	July 2013 - till date
Description	Chrysler’s learning management system which converged with FIAT’s learning management system felt a requirement for reducing number of Business Intelligence report and come up with a very less number of reports and two dashboards. Though dashboards are for summarized information, Chrysler wants a dashboard where they can select multiple filters and create their own version of detail record. This will reduce usage of complex reporting tool for Adhoc report creation and reduce number of Adhoc reports. Also they require a highly summarized executive dashboard which will connect several systems along with SABA learning management system and get the analytical values on different KPI.
Role	Business Intelligence Analyst and Designer.
Solution Environment	Windows 7
Tools	DB2, Qlikview, Microsoft Visio
Highlights	This project is a part of other LMS projects which is going on and I was assigned the analysis and design of these two dashboards. My primary responsibilities are: 1.       Understanding Business requirement 2.       Creating Mock-ups (wireframes) for visualization effects to customer.  3.       Feasibility analysis 4.       Creation of BRD and get a sign off. 5.       GAP analysis on what exist and what new data items required in data mart. 6.       Designing QVD file with data modeler for reporting. 7.       Coordinating with Qlikview developers. 8.       Testing of the data. 9.       As the project is still in requirement analysis phase my current work is maintaining coordination with Business and Modeler.

Chrysler existing system had more than 3500 reports to analysis. Primary requirement for this project was to deliver only the reports but while analysis I came to understand that there is a strong possibilities to create multiple dashboard from the reports.

                I first consolidated the list and categorize by geography. That reduces a big number. Previously there was several country specific reports which I clubbed into one and put country filter into it. Then I start looking on basis of time dimension. I came to understand that there are several reports which have no difference except the reports ran on particular time frame and saved frozen for future use. I clubbed those altogether and with modeller we design the mart which can help user to run in any time format and the history data will be visible. We had a tough time to choose between snapshotting and slow changing dimension. Later on we concluded on slow changing dimension.

                Once above two big problem was solved we remained with half of the report. Then I started analysing holistically that which reports looks similar. I started with similar in name and some functionality. I categorized a big set of report and with help from SMEs I get to know which are the important report (MoSCoW analysis) and can be used as base. So I took one report as base and try to add uncommon data item in different reports. Some merger was huge and one report consumes around 32 reports at whole.

                I end up with 170 unique functionality reports and 300+ reports which I thought as potential dashboard report. Dashboard in definition is highly summarized data modelled and developed with high visualization. But I came up with a dashboard which is detail in nature. If we club those all reports and create a multi tab huge report, it will not help them as they need an environment where they need to slice and dice data by selecting different filter, not using tools. And also the new designed dashboard is common for all departments and you cannot run a report for all departments in a single refresh.

                I created the mock up and confirmed from business and they are really happy to not use any BI tool report development functionality for their Adhoc reporting. All the filters will be present in left side of the window and in right side the report table will dynamically change with the selection. It is a cool concept of data discovery using analytics. We successfully implemented a proof of concept in Cognos and Qlikview and selected Qlikview as a tool for this dashboard. Mart level development is going on.

                From all of the reports and with the help of one enthusiastic KPI person I designed another dashboard which matches the definition of dashboard. This dashboard I am still about to analyse and complete further. As of now the dashboard mock up is completed and I am analysing the several source of data which will come in this dashboard from various application which are not part of learning managements system.