At Georgia Tech, we believe that business intelligence and analytics are all about data and governance. Simply put, it is about providing the right information to the right users at the right time. To do this, we need a strong partnership with our data stewards to ensure we are addressing the proper requirements and using the right data to do so. Creating success is an iterative process with the following key guiding principles:
- Focus on the right requirements/metrics
- Ensure a shared partnership between the Data Stewards, OIT, and EDM
- Focus on the end user
- Build a data warehouse that is a single source of integrated data at Georgia Tech. The data warehouse should house data from systems of records on campus, but should not be the system of record for any data.
- Use common/master data (ex. Chart of Accounts, People) across systems
- Start and grow responsibly and sustainably
When EDM engages with a Data Steward or a functional unit, we will follow a defined methodology to ensure the project moves smoothly through the different phases and all participants know and understand the start and ending processes.
In the Initiation Phase, EDM will work with the Data Steward to define “where are we going” and “how do we get there.” This includes the timeline, project plan, high level scope, and resources. EDM will use an estimation model to ensure Data Stewards understand the time commitments from both EDM and the functional teams.
During Requirements Gathering, our reporting process begins with identifying report requirements. To gather a full list of reporting requirements, the following questions will be used
- What reports are currently used?
- What reports are needed to meet regulatory requirements?
- What reports are needed as part of business processes?
- What system and/or business process validation reports are needed?
- What key business questions need to be answered?
- Are there any requirements that require legacy data?
Through asking these key questions, EDM will develop a comprehensive list of all reporting requirements. With the help the Data Steward(s), the reports will be prioritized and approved for build.
For each reporting requirement, a functional design specification will be created to determine how the report should look, function, and the necessary source data requirements.
Once the functional design specifications are complete for reports, all source data will be analyzed to design the data warehouse. Facts (measures) and dimensions (attributes) are identified. The data warehouse will be designed to meet reporting requirements and impose business processes on to the data. The technical design of the data warehouse consists of Source to Target mapping (source system to data warehouse) and a logical model diagram of the data warehouse.
The development of the data warehouse consists of the creation of the tables and the development of ETL (Extract, Transform, and Load) jobs. The first step of ETL jobs is to copy the data from the source systems into the data warehouse. The second step is to transform the data from the staging area into the facts and dimensions.
The reports will be developed by EDM staff using standards to ensure a common look and feel across all GT reports.
During the testing phase, the functional teams will test the reports back to the source data. This testing will ensure that not only the reports, but the data warehouse is functioning as designed
During the Deployment phase, the data warehouse tables, ETL jobs, security roles, and reports will be migrated to the production environment. EDM will work with the Data Stewards to communicate and/or train end users.
EDM is expected to provide support to end users in using the reports. We will also work with the Data Stewards to monitor key issues/enhancement requests to determine if additional changes should be made to the reports.
Our Data Warehouse team will monitor nightly ETL loads to ensure timely and accurate data for end users. Any issues with the nightly load will be communicated to end users and the Data Steward(s).