Skip to content

End-to-End Examples

This section contains complete, practical GraphReduce examples that can be run from start to finish.

These examples are intended to show:

  • dataset setup
  • graph modeling choices
  • feature and label generation
  • output inspection
  • production-oriented patterns

Start with Hello World.

Next: Preserve Child Grain (reduce=False).

Then: Multi-Backend Same Graph (pandas, sqlite, duckdb, pyspark, snowflake SQL, databricks SQL).

Then: Custom PySpark Graph (All cust_data Nodes).

Then: Custom Pandas Graph (All cust_data Nodes).

Then: Custom DuckDB Graph (All cust_data Nodes).

Then: All Tables, Full Reduction, With Labels.

Then: End-to-End Predictive AI with XGBoost.

Then: End-to-End Predictive AI with TabPFN.

Then: RelBench Rel-Stack DuckDB (Badges Label).

Then: RelBench Rel-Stack DuckDB (Post Votes Label).

Then: RelBench Rel-Stack DuckDB (User Engagement).

Then: RelBench rel-trial DuckDB (Study Outcome).

Then: RelBench rel-hm DuckDB (User Churn).

Then: RelBench rel-hm DuckDB (Item Sales).

Then: RelBench rel-avito DuckDB (User Clicks).

Then: RelBench rel-avito DuckDB (User Visits).

Then: RelBench rel-avito DuckDB (Ad CTR Regression).

Then: RelBench rel-amazon DuckDB (User Churn).

Then: RelBench rel-amazon DuckDB (Item Churn).

Then: RelBench rel-amazon DuckDB (User LTV).

Then: RelBench rel-amazon DuckDB (Item LTV).

SQL focus moved to Tutorial: SQL Backends and Dynamic Query Construction.