All Tools
B
Dev ToolsFreeOpen Source
BOKEH
Interactive visualizations for Python in the browser
BSD-3-Clause
ABOUT
ML practitioners need custom dashboards and interactive plots to visualize training metrics, model outputs, and data distributions, but building web-based UIs typically requires frontend development skills. Bokeh lets data scientists create interactive, browser-ready visualizations directly from Python — with support for linked panning, streaming data updates, and server-backed dashboards — without writing HTML or JavaScript.
INSTALL
pip install bokehINTEGRATION GUIDE
1. Build real-time training dashboards that stream loss curves, accuracy metrics, and gradient statistics during model runs
2. Create interactive model evaluation reports with linked scatter plots, histograms, and confusion matrices for stakeholder review
3. Develop custom monitoring UIs for deployed models that display prediction distributions, drift metrics, and data quality alerts
4. Embed interactive visualizations into Jupyter notebooks and standalone web applications for exploratory data analysis
TAGS
visualizationdashboardinteractivepythonplottingstreaming