Repository navigation
This release introduces a lightweight NLP-based tool for analyzing communication patterns over time.
Key features:
- Message-level labeling based on text content
- Person-level risk estimation using aggregated language patterns
- Time and event-aware analysis (e.g. late-night messages, special dates)
- Interactive Jupyter UI for adding, viewing, searching, and deleting messages
- Local SQLite storage (no cloud, no uploads)
This project is designed for reflection, not judgment.
Results are probabilistic and pattern-based — not a truth detector.
If you’re exploring mixed signals, emotional inconsistency, or communication dynamics,
this tool offers a structured way to look at the data behind the feelings.
Feedback and ideas are welcome.