Filter
Exclude
Time range
-
Near
Been trying to reproduce datashader functionality using only Apache Arrow and Acero as a learning exercise. This is a 1-billion point Clifford attractor rendered in ~8s from a 9GB parquet file. (No JIT)
1
9
655
Replying to @codetaur
It gets oversaturated when zoomed out. Are you familiar with datashader? It's Python's lib but worth checking out. It also supports interactive rendering of large datasets (hundreds of millions to billions)
2
247
Mar 12
Replying to @luigifcruz
next time try datashader with colorcet colormaps and perhaps histogram equalization for that extra punch
8
1,439
Mar 7
Replying to @bookdepth
for really heavy charts i loved datashader, the picture in my profile is made with it, it visualizes over 2 billion rows. there is some preprocessing involved, but i think the final batch i sent to datashader was still in the order of millions.
1
1
191
22 Dec 2025
Colorcet is an open-source Python library of perceptually accurate colormaps for data visualization. It offers 256-color continuous and categorical palettes that make patterns easier to see and interpret. Works with @matplotlib , @bokeh , @HoloViews , @datashader & more πŸ‘©β€πŸ”¬πŸŽ¨
1
11
4,625
23 Nov 2025
Replying to @afozsn @maruushae
working with large datasets is a breeze with polars parquet. it's surprisingly scalable, worked fine for me at a scale of billions of rows and around 1tb of data. for charts the data needs to be preprocessed anyway, i try to keep it below 10k rows. for very dense charts (think 1m rows) use datashader, i posted about this on my blog.
2
5
1,340
17 Nov 2025
Day 17 of the #30DayMapChallenge - New tool. It has been on my radar for a while so I tried out @datashader to visualise population density. These maps usually take minutes to render but with datashader it takes seconds.
3
13
91
5,943
17 Nov 2025
#30DayMapChallenge Day 17: A new tool I look into Datashader and Bokeh to process and visualize geospatial data. Here I visualize Chesapeake Bay with triangular data mesh. I rewrite this example gallery code from Datashader web: lnkd.in/gDJ7QEa9 colab.research.google.com/dr…
1
1
120
Huge props to @leland_mcinnes for optimizing the hammer bundling code in @datashader and for @BarrettLyon's inspring work at the Opte Project.
5
186
29 Sep 2025
πŸ”₯ @datashader
526.9 million player deaths in 24.7 million levels of Super Mario Maker 2. Data by @tgr_code.
3
335
28 Sep 2025
Want to visualize billions of points interactively β€” without losing detail? πŸ“Š Our sister library @datashader turns massive datasets into images, revealing true patterns & distributions in seconds. πŸ”— Link to guide below #Python #DataViz #BigData #Datashader
2
4
5
374
22 Sep 2025
🎨 Need better colormaps for your Python visualizations? Try our sister library Colorcet β€” a collection of perceptually accurate 256-color colormaps for @bokeh , @matplotlib , @HoloViews , @datashader & more πŸ”— colorcet.holoviz.org/ #python #dataviz #datascience #Analytics
2
12
3,537
That's not far off from what you described since Toponymy uses (a slightly custom variation of) HDBSCAN for clustering, and DataMapPlot uses Datashader for visualization (for static plots with enough points). They just provide extra tooling knitting more of the workflow together.
1
3
199
In this hands-on course, learn how to build fast, responsive interactive visualizations for large datasets using GPU-accelerated Python libraries: βœ”οΈ cuDF βœ”οΈ Datashader βœ”οΈ hvPlot βœ”οΈ cuXfilter βœ”οΈ Plotly Dash Perfect for developers, data scientists, and analysts working with complex or real-time data. πŸ’₯ Free for a limited time β†’ nvda.ws/4emoy0T

1
4
37
2,908
Replying to @NLP_nerd
Agreed -- always visualize first and often! A tool like @datashader that can handle all the raw data quickly and easily is great for that.
2
31
29 May 2025
Our sister library @datashader in action. πŸ‘‰ aetperf.github.io/2025/05/23…
5
216
10 Apr 2025
We've released version 0.18.0 of our sister Big Data visualization library @datashader.
3
8
297
21 Mar 2025
for high density heatmaps i highly recommend datashader
1
1
204
6 Feb 2025
umap.plot supports 'hammer' edge bundling and IIUC is internally handled with datashader. can get prohibitive on huge graphs, but runs well at this scale. very informative once you get used to it - high order bit is generally when edges are missing. umap-learn.readthedocs.io/en…

1
2
79