Access to HR information

Search dashboards for human rights case law, including the Polish Constitutional Court and the ECtHR

This project develops open-source, fully static dashboards that make human rights case law searchable at the paragraph level. Each implementation targets a different jurisdiction but shares the same core approach: structured full-text search, advanced filtering, analytics, and GitHub Pages deployment with no server required.

Polish Constitutional Court

A dashboard for exploring the case law of the Polish Constitutional Court (Trybunał Konstytucyjny), covering 3,200+ cases and 265,000+ paragraphs from 1997 to 2026.

Features

  • Paragraph-level search with instant highlighting across the full corpus.
  • Advanced filters by section (sentencja, uzasadnienie), year, decision type, reporting judge, and case signature.
  • Two modes: Student (guided, built-in dataset) and Expert (custom corpus upload).
  • Analytics page with statistics on cases by year, decision type, and judge.

Dataset: Hugging Face | Source code: GitHub

European Court of Human Rights

A dashboard for ECtHR case law with paragraph-level structural search, citation tracking, and analytics. Paragraphs are rebuilt from HUDOC source DOCX files for source-exact text fidelity. French-language judgments are included.

Features

  • Paragraph-level search across ECHR judgments with instant highlighting, filters for articles, respondent states, judges, importance level, and document type.
  • Citation extraction — each case shows which ECtHR decisions and domestic-law sources it cites, and which later cases cite it back.
  • Source-exact paragraphs rebuilt from HUDOC DOCX files, with HUDOC-faithful modal rendering including multi-column appendix tables.
  • In-browser ML classifier for creating custom paragraph labels using TF-IDF, character n-gram, or keyword-overlap models.
  • Analytics page with case statistics and visualizations.
  • Automated data pipeline with GitHub Actions for building and deploying updated datasets.

Source code: GitHub

Shared approach

  • Fully static — runs entirely in the browser, hosted on GitHub Pages.
  • JavaScript frontend with Web Workers for client-side indexing.
  • Python/Node.js data pipelines for building and normalizing datasets.
  • Open-source under permissive licenses.