Datashader

LOVE this! @Instacart delivery visualizations using #datashader from @jeremystan: bit.ly/2tkdHNY #OpenSource #DataViz #Instacart

We’re excited to announce our support for Pandata, the scalable open-source analysis stack 🎉 🐍 🌐 Pandata offers a set of general-purpose & powerful open-source data analytics tools enabling scalable data processing for any domain. Read more here 👇 bit.ly/45q2sFi

1 billion points of NYC taxi data, interactively visualized with @datashader

📢Videos from @europython are out. 💁‍♀️If you are interested in big data visualization and data apps check out Jean-Luc Stevens talk 🧵👇 Link below #python #dataviz #datascience @datashader

Code Dataset Visualization—11.66 million files from the Stack, a @BigCodeProject dataset sourced from permissively-licensed @github repositories spanning 86 programming languages (StarCoder languages subset).

Our project family continues to grow 🪴🙌! We're excited to welcome @HoloViz_org, who has joined NumFOCUS as a sponsored project! Learn more about the project here: numfocus.org/project/holoviz. #openscience #support #scientificcomputing #datascience

🎉 We’re thrilled to announce @AnacondaInc’s first ever #DataAppContest !! The competition goes from June 13-30, so start building your team for the chance to win great prizes & showcase your skills! Learn more & register here 👇 bit.ly/3VUzDxd

We are thrilled to announce the availability of the Panel 1.0 release candidate for general testing! This release represents a leap forward in performance and usability. We need your help to make this a smooth transition, find the announcement here: discourse.holoviz.org/t/panel-1-0-re…

We revamped our Bitcoin OTC user ratings example to use Hiveplotlib's @datashader capabilities. We see changes in trust in the network over time that appear to correlate with the price of Bitcoin. For more, check out our docs: geomdata.gitlab.io/hiveplotlib/bi…

Check out this new review of Python viz libraries for geographic applications, comparing Altair, @Bokeh, Cartopy, @Datashader, GeoPandas, Geoplot, GeoViews, hvPlot, and Plotly: gregorhd.github.io/geospatial-vis…

Video visualizing 1,000,000,000 points youtu.be/LKIRAzsqLb0 This is a common example with Datashader. This video is different in that it focuses on performance tuning. When we start, an update takes 40s. When we're done, it takes ~1s. High performance requires thought.

Cool. Everyone be sure to zoom in and see the pretty snowflakes when these patterns are rendered at high resolution like this...

We are excited for @HoloViews, hvPlot, and @Panel_org to start building on the upcoming better support for bioscience and biomedical applications in @bokeh !

@PratapVardhan @sujeetomar @streamlit Datashader easily handles 10 million or 500 million, and takes all data points into account when constructing the visualization (accumulating them per pixel and faithfully reporting the density per pixel). It's a good choice for anything over tens of thousands of points.

@sujeetomar @streamlit I suspect the issue here is not about streamlit. 10M points will be a bottleneck in any rendering system. SVG will not work beyond few thousand points. Canvas will take time to render million points. Do you need to plot all points, does sampling work? Our eyes only read so much.

I'll be building few data apps in @streamlit this month. This should be fun. 🐍 What's the most fun / complex / useful streamlit app you've seen/used?

Great new @Datashader talk about the UK Census from github.com/ianthomas23: youtube.com/watch?v=GX5IQg…

New post : Download some benchmark road networks for Shortest Paths algorithms bit.ly/3R5GSOv #Python #networks #Parquet @DuckDB @Datashader

@HoloViews @sophiamyang Exciting hvPlot 0.8 news: In addition to @bokeh, you can now create @matplotlib and @plotlygraphs: blog.holoviz.org/hvplot_0.8.0.h… Compared to the default Matplotlib Pandas .plot(), hvPlot adds automatic widgets for exploring data, easy layouts, maps, datashader, and more!

We are very excited to announce the release of hvPlot 0.8! Discover the new features here: blog.holoviz.org/hvplot_0.8.0.h… or check out @sophiamyang's amazing video: youtube.com/watch?v=UJtmo4…

The beauty of matrix eigenvalues. Eigen Values of 1M random matrices of form [A 1 -1; -1 B 0; 1 -1 -1] Made in #Python using @datashader and @matplotlib Inspired from the same viz by @BohemianMatrix

@mathieubastian @nicolasbchb @Neolectron @sigmajs @Gephi Datashader directly supports points, lines, areas, rasters, trimeshes (triangle-based meshes), quadmeshes (irregular rasters), and polygons, and other libraries use those primitives to support many other types of plots (see #supported-elements" target="_blank" rel="nofollow noopener">holoviews.org/user_guide/Lar…).

@datashader @nicolasbchb @Neolectron @sigmajs @Gephi I like the idea, do you guys also support edges or just points?

En exclusivité, la cartographie Twitch du #Zevent22 dans son état "primordial", avant spatialisation 🪐

@nicolasbchb @Neolectron @sigmajs @Gephi Datashader dessine des millions de points en millisecondes (ou des milliards en secondes) sur un ordinateur portable. Et si vous l'utilisez à partir de hvPlot.holoviz.org, il redessinera lors du zoom et du pan, ce qui facilitera l'exploration de "big data".

@Neolectron Alors c'est techniquement faisable et ce genre de technos existe (cc @sigmajs)... Mais pas pour des cartos de cette taille ! Sur Sigma y'a un cap de perf à ~50k nodes, j'en ai 2 millions mdr Même sur @Gephi avec des ressources de fou je peux pas naviguer dedans en temps réel