High Performance Visualization | Parallel performance with Dask & Datashader

Interactively visualize a billion points with Dask and Datashader.
This takes an example from Anaconda (thanks @Peter Wang!) visualizing the NYC Taxi data with Dask. When we start it takes a minute to render the dataset. Then, after doing a bit of performance tuning we make this fast enough to render interactively (about a second per update). This enables us to interact live with a large dataset.
Originally given as part of the Dask tutorial at PyData NYC 2022
Materials at github.com/mrocklin/dask-tuto...
More reading here:
Datashader Data Visualization Performance
www.coiled.io/blog/datashader...
---
Scale Your Python Workloads with Dask and Coiled.
Coiled is a Dask company. With Coiled's rock-solid infrastructure, you can quickly and securely create Dask clusters in your cloud account.
Learn more about Coiled and get started for free
coiled.io/start
More content on our blog:
coiled.io/blog

[@floopybits8037]

Just one word WOW

[@gatorpika]

Very cool, thanks for this! I have been playing around with datashader a bit on my laptop and it renders tens of millions of records pretty well, but then I run out of memory. We are looking to move stuff to the cloud and this will be helpful in using that to visualize larger sets of IOT data.

[@billyblackburn864]

the one at 15min is really nice...what is the cluster you're running it on?