How I mapped every High Court of Australia case and their citations (1901-2025)
A new legal AI project visualizes 7,000+ cases and 20,000 citations with semantic embeddings and sentiment analysis.
A developer has built a comprehensive legal knowledge graph visualizing 124 years of Australian High Court jurisprudence. Using the Open Australian Legal Corpus, they processed approximately 7,000 cases spanning 1901-2025, extracting 20,000+ citation relationships between High Court decisions. The project employed Kanon 2's AI tools for citation extraction and embedding generation, creating vector representations of each case's semantic content. Through dimensionality reduction (PaCMAP) and clustering (K-means), cases were mapped into 3D space where proximity indicates legal similarity. The system also analyzes citation sentiment, color-coding supportive (green) versus overruling (red) references. This creates an unprecedented interactive visualization where users can explore how different areas of law (criminal, estate, land law) cluster together and evolve over time, revealing patterns invisible in traditional legal research.
- Processed 7,000+ High Court cases from 1901-2025 with 20,000+ citation relationships using Kanon 2 AI tools
- Generated semantic embeddings and reduced to 3D visualizations showing how related legal areas cluster together
- Added sentiment analysis to citations (green=supportive, red=overruling) creating unprecedented legal relationship mapping
Why It Matters
Demonstrates how AI can transform legal research by making complex citation networks and legal evolution visually accessible and data-driven.