Questions that could potentially be answered by the dataviz evaluation project

arviz-devs / GSoD

Repo about ArviZ participation in Google Season of Docs initiative

Apache License 2.0

5 stars 2 forks source link

Questions that could potentially be answered by the dataviz evaluation project #6

Open OriolAbril opened 2 years ago

OriolAbril commented 2 years ago

Defaults for 2d kdes and plot_pair -> scatter vs kde vs hexbin.
- Should it depend on the number of samples to plot?
- Or should the default always be the same but some types like scatter should be discouraged if we have >10k datapoints?
- For 2d kde in general, is the density levels default good? Or should the default be levels defined by the probability within them maybe also with a label in the contour line?

OriolAbril commented 2 years ago

We could also add a plot subtitle with a short sentence on how to interpret plots. i.e. https://easystats.github.io/performance/index.html#comprehensive-visualization-of-model-checks

aloctavodia commented 2 years ago

What about using logging for this and a rcparams to control de verbosity?

OriolAbril commented 2 years ago

being an rcParam would be great. Not sure about logging because I'd expect it is more common for the audience to need the subtitle, not the person who actually generates the plot.