Database structure

Database structure

The CBS Analysis Database is designed and implemented as a Data Warehouse. Data Warehouse databases are the core component of Business Intelligence, integrating data from various sources into a dimensional model. This dimensional model facilitates aggregation, and slicing and dicing of the data. It uses the concept of facts and dimensions. Events are translated into ‘facts’ and the actors in the event (e.g. the client, the facility, the geographical location) are the ‘dimensions’. An example of the sentinel event “ART initiation” in the CBS database:

The CBS database is populated with the data from the various sources as described in the Data sources page using a technique called ETL (Extract, Transfer, Load). This is a step-wise approach in which the data is first extracted from the source systems, thereafter transformed into the dimensional model, and lastly loaded into the CBS Analysis Database. The transformation includes data cleaning, performing calculations and derivations, and finally the re-modelling into the dimensions and facts.

In the ETL the sentinel events are identified and key indicators calculated. More about the key indicators in the Data Analysis section.