SHAP Explorer — Explain Your Model

Train a small Random Forest and explore global and local SHAP-style explanations. See which features matter overall and why a specific prediction happened.

Global Local Guide
Beginner Quick Start
  1. Select a dataset (regression/classification) and click Generate.
  2. Train the Random Forest (defaults work well).
  3. Open Global to see which features matter; open Local to explain a specific row.
Tip: In Local, try the what‑if sliders to see how changing a feature affects the prediction and attributions.
Global Explanations
Global importance = mean |SHAP| per feature (top‑k).
Per‑feature SHAP “strip” (beeswarm‑style): x=SHAP value, y=feature; color = normalized feature value.
Local Explanation (Waterfall)
Base value + contributions → prediction
FeatureValueSHAP
Edit selected row (numeric features)
Dataset
Rows600
Train %80%
Model (Random Forest)
Trees100
Max depth5
Min leaf3
Metric: — —
SHAP & Instance
Top‑k (global/local)10
What is SHAP — and how to read these plots
SHAP in one minute
  • SHAP (Shapley Additive Explanations) attributes a model prediction to features fairly, inspired by Shapley values from game theory.
  • Local: Explain one prediction (waterfall shows how each feature pushes it ↑/↓ from the base value).
  • Global: Summarize which features matter overall (mean |SHAP|) and distribution by feature (strip/beeswarm).
Reading the plots
  • Global bar: Taller = more important on average.
  • Strip: Spread shows variability; sign shows direction; color hints at feature value trends.
  • Waterfall: Positive bar pushes prediction up; negative down. Sum + base value = prediction.
Good to know
  • These attributions are explanations, not causality.
  • Correlated features can share credit; examine dependence trends.
  • Faster method used here: path-based contributions on trees (TreeSHAP‑inspired) for snappy interaction.
Next steps: KernelSHAP option, real datasets, multiclass views, and richer dependence plots.

Support This Free Tool

Every coffee helps keep the servers running. Every book sale funds the next tool I'm dreaming up. You're not just supporting a site — you're helping me build what developers actually need.

500K+ users
200+ tools
100% private
☕ One-time support Buy me a coffee 📚 Learn & support 9-Book Bundle - $9 Stay updated Follow @anish2good
Privacy Guarantee: Private keys you enter or generate are never stored on our servers. All tools are served over HTTPS.

About This Tool & Methodology

SHAP (SHapley Additive exPlanations) attributes a model’s prediction to each feature using game‑theoretic Shapley values. This explorer visualizes local (per‑sample) and global (aggregate) attributions and supports dependence/summary plots. For performance, simplified background distributions may be used.

Learning Outcomes

  • Interpret SHAP values as feature contributions to individual predictions.
  • Distinguish global importance vs local explanations.
  • Recognize interactions and correlated features caveats.

Authorship & Review

  • Author: 8gwifi.org engineering team
  • Reviewed by: Anish Nath
  • Last updated: 2025-11-19

Trust & Privacy

  • Runs locally with sample or provided data; no uploads are stored.
Quick Access