holoviz/datashader

Health v2 · maintenance only

Adoption not part of the score

3558 stars · 376 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Datashader is a Python data rasterization pipeline that renders very large datasets into fixed-size images by projecting, aggregating, and shading data into grids. It is part of the HoloViz ecosystem and uses Numba, Dask, and CUDA to scale across CPU cores and GPUs.

Use cases

• plot millions of data points without subsampling
• render large scatter plots of geographic coordinates
• visualize distributions of huge datasets as images
• rasterize large time series or line charts
• create density heatmaps from big data
• visualize large datasets in a browser dashboard

When to choose

• your dataset is too large for conventional plotting libraries and you are subsampling or chunking just to plot it
• you need faithful, accurate representation of the full data distribution
• you want GPU or distributed rendering of large data
• you are already using HoloViz tools like HoloViews, hvPlot, or Panel

When to avoid

• you have only hundreds or thousands of points where standard tools like Bokeh or Matplotlib suffice
• you need per-point interactivity, metadata, or hover on individual data points
• you need a CLI tool (the CLI was removed in 0.19.0)

Facets

library · maturity active

data-visualization image-processing gpu-computing data-visualization data-science big-data python cross-platform rasterization numba dask holoviz large-datasets plotting geospatial gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="holoviz/datashader")