If you create visualizations with #Python, what would be your 2026 choice?
-
Tagging my favorite accounts for data viz @Luca and @kristinHenry
-
@guerda I'd always go with matplotlib due to its flexibility. You can customise absolutely everything! But then sometimes I mix in a bit of seaborn for specific chart types it does better than native matplotlib, eg heatmaps or stripplots.
-
@guerda I'd always go with matplotlib due to its flexibility. You can customise absolutely everything! But then sometimes I mix in a bit of seaborn for specific chart types it does better than native matplotlib, eg heatmaps or stripplots.
@LisaHornung I agree with matplotlib, I often fall back on it to configure specific aspects of plots. I still find the interface quite confusing still (subplot, figure, axes, ...)
-
@guerda Love mpl for flexibility, but nowadays mostly use plotly for interactivity.
-
@guerda Love mpl for flexibility, but nowadays mostly use plotly for interactivity.
@inviridi yeah, plotly is great, I will check it out even though I don't need the interactivity for this case..
-
Any experience with #bokeh, #plotnine or #VegaAltair?
-
@inviridi yeah, plotly is great, I will check it out even though I don't need the interactivity for this case..
@guerda Things often start out simple, and at one point you just realize that you need the big machinery to get the job done. That is matplotlib. But if you do not reach that point, plotly is awesome. One of the things I like most about it, is that all properties are simple dicts -- and JSON. So if you made a nice chart, it works in JavaScript as well.
-
F flumen_calculi@ruhr.social shared this topic
