# MAIF/shapash

🔅 Shapash: User-friendly Explainability and Interpretability to Develop Reliable and Transparent Machine Learning Models

Repository: https://github.com/MAIF/shapash
Canonical: https://ross.abutalabs.com/products/shapash
Homepage: https://maif.github.io/shapash/
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Topics: python, machine-learning, explainability, explainable-ml, transparency, ethical-artificial-intelligence, shap, lime, interpretability
Last push: 2026-08-25T12:28:53+00:00

## Health v2 (maintenance only)
Score: 90/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 74, longevity 100
- inputs: {"age_days": 2317, "days_push": 8, "days_rel": 98, "gap_med": 31.5, "n_releases_24m": 13}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3251, forks 387 (observed 2026-08-28T04:07:52.120690+00:00)

## What it is
Shapash is a Python library that makes machine learning models interpretable and understandable through clear visualizations, a webapp, and audit reports. It builds on SHAP and LIME backends to provide global and local explainability for models like CatBoost, XGBoost, LightGBM, and scikit-learn.

## Use cases
- explain machine learning model predictions to non-technical stakeholders
- visualize global and local feature importance
- generate an audit report for a data science project
- build a webapp to explore model explanations
- summarize local explanations for end users
- check quality of explainability methods
- interpret xgboost or lightgbm model outputs

## When to choose
- you need user-friendly, labeled visualizations of SHAP/LIME explanations
- you want to share model insights with both data scientists and business users
- you need documentation or audit reports for model transparency
- you use common Python ML stacks like sklearn, XGBoost, LightGBM, or CatBoost

## When to avoid
- you need explainability outside the Python ecosystem
- your model is not among the supported frameworks and you cannot provide a custom backend
- you only need raw SHAP values without summarization or visualization

## Facets
- artifact type: library
- maturity: active
- function: data-visualization, machine-learning, nlp
- domain: machine-learning, data-science, data-visualization, artificial-intelligence
- platform: python
- tags: explainability, interpretability, shap, lime, xai, model-auditing, webapp, ethical-ai

## Member repositories
- MAIF/shapash (main) score 90

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:52.120690+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T07:24:25.087803+00:00, confidence not recorded.
  - readme: https://github.com/MAIF/shapash (fetched 2026-08-28T04:07:52.120690+00:00, sha fb85c6fa3852)
  - homepage: https://maif.github.io/shapash/ (fetched 2026-08-29T09:37:00.670191+00:00, sha 1b3e6de27d79)
  - registry_pypi: https://pypi.org/pypi/shapash/json (fetched 2026-08-29T09:37:00.696454+00:00, sha e8e43a94684f)
- Data as of 2026-08-30T08:39:29.467469+00:00.