Tools overview

Beyond forecasting, RINK includes four additional ML / optimization tools that share the same authentication, per-user file handling, and visual language.

Tool	Best for
Forecasting	Predicting future values of a time series
Anomaly Detection	Flagging unusual rows for fraud / fault detection
Churn Prediction	Scoring which customers are likely to leave
Customer Segmentation	Clustering customers into natural groups (K-means)
A/B Test Analyzer	Statistical significance for continuous or conversion experiments
TSP	Single-vehicle shortest route through N locations
Vehicle Routing	Multi-vehicle fleet planning with capacity constraints

All tools are accessible from the Tools dropdown in the navbar once you're signed in.

Common patterns

• Inputs: most tools accept a CSV (≤ 10 MB) drag-and-dropped into the workspace. TSP and VRP accept tabular text input instead, so you can paste coordinates directly.
• Outputs: every tool renders KPIs at the top, a primary visualisation in the centre, and tool-specific detail tables / lists below.
• Privacy: uploaded data is stored under your user directory and deleted on sign-out (manual or after the 4-hour idle timeout).
• Auth: every tool's API requires a valid Supabase access token; the client attaches it automatically via the api.js interceptor.