A lot of people use PostGIS as a basic GIS toolbox, but very few use it in a real advanced way. To progress towards full PostGIS power, we can first make use of advanced native PostGIS functions. Using some extensions related to PostGIS, such as SFCGAL (for 3D data management), PostGIS Raster, PgPointCloud or even the latest pgsql-postal (for address normalization)... Then we can mix PostGIS functions with advanced standardized SQL features provided by PostgreSQL 9.x itself (CTE, Window functions, FDW, join and aggregate pushdowns…). Even better, use PostgreSQL bindings for data analysis languages such as R or Python to create your own dedicated function set, and integrate them into your SQL queries. We conclude this session with perspectives on what the combination between GIS and data science could be.