Workshop: From Entities to Edges – A Practical Introduction to Knowledge Graphs

Instructor: Laurence Brandenberger, University of Zurich Assistant: Yaren Durgun, University of Zurich

---

📘 Overview

This workshop introduces social scientists to the principles and practice of Knowledge Graphs (KGs). KGs are data structures for connecting and querying complex information.

You’ll learn how to:

• Understand what Knowledge Graphs are and why they matter.
• Build a simple Neo4j Knowledge Graph from CSV or Excel data.
• Validate, query, and explore your graph using Cypher.
• Apply these skills to real-world research data.

---

🗂️ Repository Structure

Workshop_KnowledgeGraph_for_SocialScientists/
│
├── Tutorial1_Simple/             # Example 1: Northwind business dataset
│   ├── example_northwind.ipynb
│
├── Tutorial2_UKParliament/       # Example 2: UK Parliament dataset
│   ├── load_parliament.py        # Main script for data import
│   ├── parliament_modules.py     # Custom wrappers and preprocessors
│   ├── conversion_schema.yaml    # Schema defining entities and relationships
│   ├── debates.xlsx              # Source data
│   └── requirements.txt          # Python dependencies
│
├── Slides_Workshop_Introduction_to_KnowledgeGraphs.pdf  # Workshop slides (PDF)
│
└── README.md                     # This file

---

🧰 Requirements

• Python 3.10+
• Neo4j Desktop (free community edition)
• Recommended: a virtual environment (venv)

Install dependencies:

rm -rf .venv
python3.12 -m venv .venv
source .venv/bin/activate
pip install -r Tutorial2_UKParliament/requirements.txt

---

🚀 How to Run the Tutorials

1. Northwind Example

A clean, minimal example for understanding how data2neo works.

jupyter notebook Tutorial1_Simple/example_northwind.ipynb

2. UK Parliament Example

A realistic, messy dataset to practice schema design, wrappers, and validation.

cd Tutorial2_UKParliament
python load_parliament.py

After import, open Neo4j Browser and explore:

CALL db.schema.visualization();

---

🧠 Learning Outcomes

By the end of this workshop, you will be able to:

• Explain the difference between relational and graph-based data.
• Construct a simple knowledge graph using Neo4j and data2neo.
• Write Cypher queries to explore entities, relationships, and attributes.
• Identify how KGs can improve research transparency and reproducibility.

---

🔗 Resources

• Neo4j Desktop Download
• data2neo Github
• data2neo Documentation
• Neo4j Cypher Manual

---

🧩 License

This material is provided for educational and research purposes under the MIT License.

---

🙌 Acknowledgements

Developed as part of the DemocraSci Project at the University of Zurich.
Special thanks to Julian Minder for the data2neo library.