ExplainerDashboard is an open-source Python library that turns any machine-learning model into an interactive Explainable AI dashboard in minutes, showing exactly how it makes predictions. It has around 2,500 GitHub stars.
Built by Oege Dijk, the library wraps scikit-learn-style classifiers and regressors in a web dashboard built on Dash and Plotly. With a few lines of code you get SHAP values, permutation importances, partial dependence plots, individual prediction breakdowns and what-if analysis, all explorable in the browser. No frontend code is required; the library generates the full app from your fitted model and data. The dashboard covers model-level questions โ which features drive predictions overall, where performance breaks down โ and instance-level ones, such as why one specific customer was scored the way they were. Components are modular, so you can assemble custom dashboards from individual plots, and the same explainer objects work in notebooks for quick analysis. It is a favourite for regulated industries and data-science teams who must justify model behaviour to non-technical stakeholders. The project is actively maintained, with updates in 2026.
ExplainerDashboard is free and open source under the MIT license. Dashboards run on your own infrastructure at no cost; deploy it behind your own authentication like any other Dash app.
Data scientists, ML engineers and risk teams who need to explain, audit and debug black-box models for stakeholders or regulators, and anyone teaching explainable AI concepts.