I used NLP and behavioral tagging to visualize abuse escalation patterns over time — here’s what that looks like

I’m a behavior analyst and trauma researcher building a project called Tether, which uses a multi-label NLP model to tag abusive language patterns (e.g., gaslighting, control, DARVO, threats). One of the most powerful features we’ve developed is a timeline visualization that maps escalation patterns in real relationships over time.

🧠 Each message is labeled by abuse type, emotional tone, behavior function, and escalation risk.

📈 The data is then used to generate plots showing:

• Abuse intensity over time
• DARVO probability spikes
• Emotional tone shifts (supportive vs. undermining)
• Composite risk scoring for user reflection and intervention

These charts help survivors and clinicians see what’s usually only felt.

If this kind of behavioral + language mapping interests you, I’m happy to share visuals or the app itself.

Note: The tool is not for real-time diagnosis or moderation—it’s a personal safety reflection tool grounded in behavioral science.