We talk a lot about “data-driven decisions”, but that usually hides three separate layers: Data itself (events, transactions, logs, etc.). Database structure (schemas, constraints, relationships). Insights on top (from SQL, AI copilots, BI tools, notebooks). My current interest is in that middle layer: using real-world database structures as a playground to practice database insights : Understanding and improving data quality (missing constraints, odd cardinalities, misuse of types). Suggesting schema improvements for future analytics (indexes, normalization vs denormalization, slowly changing dimensions, etc.). Making access easier via views, semantic layers, or documentation (ERDs, db docs, column descriptions). I’m looking for good open source repositories that contain: Non-trivial database schemas (preferably SQL migrations or DBML, not just ORM models). Some realistic sample data if possible (to check cardinalities, null patterns, etc.).…