Datashader: Plotting all your data
Dr. James A. Bednar, who works at Anaconda running consulting projects and maintaining the HoloViz suite of open-source tools, joins Coiled to discuss Datashader.
Datashader renders even the largest datasets quickly and accurately, helping reveal their true distribution.
Without Datashader, additional data points often obscure rather than reveal your data, with each new point overplotting and hiding previous ones. Datashader numerically aggregates all the data points falling in each pixel, then uses nonlinear colormapping to visualize both trends and outliers faithfully.
Datashader’s computation is accelerated on CPUs using Numba (+ RAPIDS on GPUs) and distributed using Dask so that you can make full use of all available hardware. To help you see patterns at every spatial scale, Datashader has been integrated with Bokeh and Plotly (via HoloViews) and with Matplotlib, providing interactive plotting for billions of points even on an ordinary laptop.
After attending, you’ll know:
- What sort of plotting issues Datashader helps you avoid
- How to apply Datashader to your Pandas or Xarray data
- When to use Dask with Datashader
Join us Thursday, July 29th at 6 pm US Eastern time by signing up here and dive into the wonderful world of Datashader!