Data models are essential for analytical purposes. A data warehouse (DWH) is used for storing historical data. This provides insights into trends over time from previous events and can be used to make predictions for the future.
9 out of 10 times, a new data model results from reverse engineering. Data vault is often used for the backend because of its scalability, flexibility, and auditability and traceability.
For the front-end, a Kimball model (dimensional or star schema) is the one of best solutions. A Kimball model is easy to understand and use, and it allows for efficient and accurate reporting.
Initially, the backend data model is built, taking as much data as possible into account. In the front-end model, a selection is made of relevant data in a star schema, on which the reports are based.