Bokeh Get started

Interactive visualization for Python

Interactive visualizations for modern browsers.

Bokeh is a Python library for interactive plots and data applications in modern web browsers. Start with high-level plotting tools, then use layouts, widgets, and models when you need more control.

pip install bokeh
  • Python-native
  • No JavaScript required
  • BSD licensed
Selected examples
Bokeh terrain contour application with a movable transect across a modeled landscape Server appTerrain contour explorer Bokeh choropleth map of Texas counties with a hover tooltip GeospatialInteractive choropleth Bokeh plot of Bessel functions with a LaTeX equation, mathematical labels, and axis notation ScientificMath and LaTeX Bokeh candlestick chart showing Microsoft stock prices Time seriesMarket series A Bokeh network graph created from a NetworkX graph Graph dataNetwork graph Bokeh server application showing an interactive population pyramid and projected trend Server appPopulation explorer

Common workflows

Use Bokeh where you already work.

Build plots in Jupyter, publish them on a web page, or add Python callbacks with the Bokeh server.

01Jupyter
A Bokeh ridge plot rendered inline below Python code in JupyterLab

Explore data in notebooks

Place interactive plots beside your analysis in Jupyter notebooks and JupyterLab.

Use Bokeh in Jupyter
02Standalone

Publish anywhere

Save a standalone HTML file or embed a plot in an existing page. Neither option needs a Python server.

Publish and embed
03ASGI
The animated Dask diagnostic dashboard built with Bokeh

Add live Python callbacks

Mount Bokeh in FastAPI, Starlette, Django, or another ASGI app, and serve it with Uvicorn or Hypercorn. No separate Bokeh process is needed.

Use Bokeh with ASGI

A small example

Python to the browser.

Bokeh handles rendering in the browser and provides tools for pan, zoom, hover, and selection.

See the first steps
first_plot.pyPython 
01 from bokeh.plotting import figure, show
02
03 x = [1, 2, 3, 4, 5]
04 y = [6, 7, 2, 4, 5]
05
06 p = figure("A first Bokeh plot")
07 p.line(x, y, line_width=3)
08 p.scatter(x, y, size=9)
09 show(p)
Output ready in your browser

Bokeh at a glance

Built for real analytical work.

Use bokeh.plotting to make a first chart. For a custom application, work directly with plots, widgets, layouts, and other Bokeh models.

A

Built-in interaction

Bokeh plots include tools for pan, zoom, hover, selection, linking, filtering, and editable data.

B

High-level and low-level APIs

Start with bokeh.plotting, then use layouts, widgets, annotations, models, themes, and extensions as needed.

C

Streaming and larger data

Stream or patch data without replacing the full data source. Supported glyphs can use WebGL acceleration.

D

Several ways to publish

Show plots in notebooks, save standalone HTML, embed components in a web app, or run a Bokeh server.

Live demos

Try Bokeh in a live application.

View all demos
Vehicle efficiency explorer with a linked scatter plot and fuel economy distribution
Linked selectionVehicle data

Compare vehicles across linked views.

Select a group in the scatter plot to update its fuel economy distribution and summary values.
Regional airport application with airport locations plotted over map tiles
GeospatialTile map

Compare access to regional airports.

Choose a state or territory, then select an airport to find its closest alternatives.
Streaming market simulator with an updating OHLC chart
Streaming dataSession state

Watch market indicators react to new bars.

Generate an intraday OHLC stream, change its market regime, and watch volatility, drawdown, and MACD update.
Terrain contour application with a movable transect across a modeled landscape
Filled contoursLinked profile

Move a transect across modeled terrain.

Position the transect to update its elevation profile and terrain statistics.
Hubble deep-field image in an interactive image-processing application
Image processingXarray + Numba

Process a Hubble image interactively.

Crop the image and apply compiled color, contrast, blur, edge, and sharpening filters.
Nonlinear oscillator application showing phase paths for a mathematical model
Applied mathematicsSimulation

Compare nonlinear oscillators.

Switch among four models, then compare their traces, phase paths, and diagnostics.

User showcase

Projects that use Bokeh.

Dask, Panel, Mistic, and ArviZ use Bokeh for browser graphics, interaction, or application infrastructure.

01Distributed computing
Dask

Monitor distributed work.

The Dask diagnostics dashboard uses Bokeh to show task progress, memory, CPU, data transfer, and worker activity.

Visit project
02Application framework
Panel

Build Python dashboards and applications.

Panel builds on Bokeh browser models and the Bokeh server, adding its own layouts, widgets, and application APIs.

Visit project
03Biomedical imaging
Mistic

Explore multiplexed tissue images.

Mistic uses Bokeh canvases and linked interactions to compare multiplex images and investigate image t-SNE embeddings.

Visit project
04Bayesian analysis
ArviZ

Inspect Bayesian model diagnostics.

ArviZ includes a Bokeh backend for interactive posterior analysis, statistical graphics, and notebooks.

Visit project

Open source

Open source visualization,
driven by community.

Bokeh is a fiscally sponsored NumFOCUS project, built and maintained by contributors on GitHub.

Join the community Contribute on GitHub