A data model is the structure in which data and its mutual relationships are captured: which tables there are (customers, orders, products), which fields they contain and how they are linked together. In BI tools such as Power BI, the data model is the layer on which all calculations and visualisations rest.
Data model in practice
The data model is what makes a dashboard reliable. If orders, customers and payments are neatly related to one another, you can ask any question — revenue per customer per month, margin per product group — without new data exports. A poor model comes back to bite you in slow reports and figures that differ ever so slightly from one chart to the next.
A rule of thumb from practice: eighty percent of the work on a good dashboard is in the data model and the data preparation, twenty percent in the visualisation. That is why the price of dashboards varies so much from case to case — see What does a Power BI dashboard cost?
Related terms
- BI — BI (business intelligence) is the collecting, combining and visualising of business data to base decisions on — usually in the form of dashboards and reports.
- Power Query — Power Query is the tool in Excel and Power BI with which you retrieve, clean and reshape data from sources — without programming.
- DAX — DAX (Data Analysis Expressions) is the formula language of Power BI (and Excel Power Pivot) with which you define calculations over the data model: revenue compared to last year, moving averages, margins per segment.
- Single source of truth — A single source of truth is the agreement that every piece of data has one leading source: customer data lives in the CRM, inventory in the point-of-sale system or WMS, invoices in the accounting.