Collection of tools and interfaces for analyzing Twitter/X data, community networks, and exploring LLM memorization of social media content.
For the current evidence-backed discovery approach, start with methods and lessons and the code/data boundary. The private research dataset is not included. Historical map scores are not calibrated community membership or a ready-made outreach audience.
Source-backed discovery and historical graph analysis using Community Archive and independently saved evidence. Includes local search, typed relationships, curator review and legacy community visualizations. Dataset size and source availability vary by the local capture.
Each subdirectory is a self-contained project with its own dependencies, documentation, and tooling. This workspace currently includes tpot-analyzer; start with its README for setup and usage.