lecture-python-intro icon indicating copy to clipboard operation
lecture-python-intro copied to clipboard

Improve Markov chain networkx figure

Open jstac opened this issue 1 year ago • 2 comments
trafficstars

In https://intro.quantecon.org/markov_chains_I.html#example-3, the directions of the arrows and the weights are a bit hard to see. I'd like to have a version where all the transition directions and weights are clearer.

jstac avatar Mar 01 '24 19:03 jstac

@jstac this is an interesting one. There is only one large transition probability which is throwing the colorbar off. Using a diverging colormap shows there are a lot of small values.

Screenshot 2024-04-03 at 2 50 26 PM

I think we want a sequential colourmap as it is an indicator of intensity. Wondering if we should drop edges that are below are certain cutoff to clean up the visualisation and push some of the colourmap variation towards higher intensity blues.

mmcky avatar Apr 03 '24 03:04 mmcky

Thanks @mmcky , this is definitely an improvement.

I think we might need to keep all edges so that we understand if it's irreducible etc.

  • For some reason, the arrows are partially hidden. Is it possible to fix this?
  • Let's have hot colors for high intensity and cool colors for low intensity.
  • It's a shame that some colors are very pale. Maybe switch to jet colormap?

jstac avatar Apr 03 '24 17:04 jstac