VRT is the Flemish public radio and television broadcaster. VRT wants to strengthen the Flemish society by informing, inspiring, and connecting. As a radio and television broadcaster, they focus on informative and cultural programs, as well as sports, drama and entertainment.
Some of VRT’s well-known brands are:
The existing financial reporting structure of VRT was partially based on manually processed data (using Excel). This sometimes resulted in different sources of the truth, each with differing versions depending on the last download. Meetings scheduled to make informed decisions based on data could get overshadowed by discussions about the quality or the accuracy of the data itself. Cloubis was contracted with a clear goal: process the financial data via the data lake without any manual steps, and connect it to Power BI to arrive at an enterprise grade reporting solution. The first objective was to recreate an existing budget versus actuals report.
A chain of SQL statements was created in dbt to transform the source data in the data lake into proper dimension and fact tables. These tables were then loaded into Power BI’s preferred star schema and leveraged by DAX to calculate metrics that work in an interactive way. Examples of such metrics include the redistribution of the yearly budget to a monthly grain based on the actuals of the previous year, or the creation of a dynamic budget which adapts itself to the actuals.
The report has different types of end users, each with their own interests:
The goal was to capture these various interests into a single report. To achieve this, we designed the report to be interactive, with multiple layers through which the user can navigate: the most detailed layer contains transactional data, other layers contain aggregation of those transactions or visualizations of those aggregations. The key performance indicators of the report were collected in a dashboard for a more general overview. This also allows for automated notification triggers if certain thresholds should be exceeded.
Each user can now work with the same data and extract information that’s relevant to them from the same report. This allows conversations to be about the numbers rather than their accuracy. The next step is to expand the current data model to allow for other specific reporting (such as P&L), or to combine it with data from other data models (such as HR or content).