Knowledge Graphs for Data Interoperability (KG4DI)

May 6, 2026 · View on GitHub

Tutorial materials for "Knowledge Graphs for Data Interoperability with Chimera (KG4DI)" — a half-day tutorial introducing participants to Chimera, an open-source framework for building declarative, composable semantic data transformation pipelines on top of Apache Camel.

📄 Tutorial website: https://cefriel.github.io/kg4di

Overview

Participants design and execute a complete data integration pipeline — from ingestion of structured data to RDF lifting, SPARQL-based enrichment, SPARQL construction, and RDF lowering — using only YAML route definitions and declarative mapping templates. No application code is required.

The running example integrates public transport stop data (GTFS format) with geographic and descriptive information from Wikidata, visualised on an interactive map.

Prerequisites

Required

  • Docker — used to run all exercises without any local JDK or Python installation.
    • Pull the Chimera image before starting: docker pull cefriel/chimera:kg4di
    • Pull the dashboard image: docker pull cefriel/chimera:kg4di-dashboard

Alternative (local execution)

  • JBang — allows running Chimera pipelines directly on your machine.
    • Install JBang before proceeding (first time only).
    • Install Apache Camel via JBang (first time only): 
      jbang app install camel@apache/camel

Repository Structure

camel-routes-exercises/
  hello-world/   # Introductory exercise: Apache Camel basics
  e0/            # Exercise 0: RDF Lifting (CSV → RDF)
  e1/            # Exercise 1: RDF Lowering (RDF → CSV)
  e2/            # Exercise 2: Full pipeline (GTFS → RDF → Wikidata enrichment → visualization)  
visualization/   # Interactive map dashboard

Exercises

Hello World

A minimal Apache Camel route — use this to verify your setup is working correctly before attempting the main exercises.

Exercise 0 — RDF Lifting

Lift a CSV file (stops.txt) into an RDF knowledge graph by completing an MTL lifting template.

Exercise 1 — RDF Lowering

Lower an RDF knowledge graph back into a CSV representation by completing an MTL lowering template.

Exercise 2 — Full Pipeline

Build a complete end-to-end pipeline: ingest GTFS data, lift it to RDF, enrich via Wikidata SPARQL CONSTRUCT, and send the result to a visualization backend.

Visualization

The visualization/ folder contains an interactive map dashboard that displays the results of the pipeline. It can be started alongside Exercise 2 using the included docker-compose.yaml at the repository root:

docker compose up

The dashboard will be available at http://localhost:8000.

Acknowledgements

This work has been partially funded by the European Union's Horizon Europe programme under grant agreements No. 101140087 (SMARTY), No. 101092908 (SmartEdge), and No. 101239472 (UrbanFlow).