Analysis of traffic through tracking of moving objects, a location-based service, is an activity that requires storing and querying an amount of data which is typically too large to use the classical OLTP approach in a practical way.
A spatiotemporal data warehouse for moving objects is built, with different designs. Observations in position (X, Y, road segment) and time for cars, utilization of roads, and length and duration of trips are stored facts in the DW. The DW is populated from two data sources, one from real, GPS data from moving cars, and one with synthetic data from a data generator. Indexes and materialized views are built upon this DW to improve performance. A set of typical queries is issued to the DW to test this performance and compare the designs.