Bokeh

What Bokeh is

Bokeh is a free, BSD-licensed Python library for creating interactive plots and data applications in modern web browsers. Viewers can pan, zoom, hover, select, link, and filter plots, and work with editable data. Developers can use high-level plotting tools or lower-level layouts, widgets, annotations, models, themes, and extensions. Bokeh accepts Python lists, NumPy arrays, ColumnDataSource objects, pandas and Polars DataFrames, PyArrow DataFrames, pandas GroupBy objects, and REST endpoints as data sources. Data can be streamed or patched without replacing the full source, and supported glyphs can use WebGL acceleration.

Plots can be displayed in notebooks, saved as standalone HTML, embedded in web applications, or served with Bokeh server. The project describes using Bokeh in Jupyter and mounting it in FastAPI, Starlette, Django, or another ASGI application served with Uvicorn or Hypercorn. It is Python-native and supports JavaScript-powered visualizations without requiring users to write JavaScript themselves. Bokeh also lists interactive dashboards, scheduled refreshes, and embedded dashboards. The project is fiscally sponsored by NumFOCUS and maintained by contributors on GitHub. Support is volunteer-based, with usage questions directed to its public Discourse forum and bug reports or feature requests to GitHub.

Who it suits

Bokeh suits Python developers who need interactive browser plots or data applications and want to publish them in notebooks, HTML, or web apps. It also fits teams working with its listed Python data formats or REST endpoints.

Getting started

1. Visit the Bokeh website.
2. Install and use the Python library in your environment.
3. Connect a supported data source, such as a DataFrame or REST endpoint.
4. Build plots with its high-level plotting tools or lower-level components.
5. Show plots in a notebook, save them as standalone HTML, embed them in a web app, or serve them with Bokeh server.

Verdict

Choose Bokeh if you want a free Python library for interactive browser plots, with several options for publishing and serving them. Its volunteer-based support model may be a consideration for teams that need a different support arrangement.