Retention Risk Pipeline

How to Use

1. Click any pipeline stage (input stream, processing step, or output category) to see a detail panel at the bottom describing the algorithms and data involved
2. Toggle Contagion button shows or hides the contagion overlay processing stage and the contagion alert output category, demonstrating how network effects amplify individual risk signals
3. Watch the data particles flow left to right through the pipeline, visualizing how information moves from raw signals through processing to actionable risk categories
4. Click Reset to clear the detail panel and return to the default view

About

This simulation visualizes the complete retention risk pipeline used in organizational analytics. Three categories of input signals -- graph metrics from the collaboration network, NLP signals from communication analysis, and behavioral events from HR systems -- merge into a feature engineering stage. The engineered features feed a machine learning prediction model that classifies employees into risk tiers: low risk (monitor quarterly), watch (monthly check-in), and high risk (immediate intervention).

The optional contagion overlay adds a network-aware layer that detects when departure risk spreads through tightly connected teams, triggering team-level action rather than just individual interventions. This reflects research showing that turnover is often contagious within close-knit organizational clusters.