Adding a legend to interactive visualizations

[fl-hat]

Hi! I've been using an ipysimulate scatterplot (similar to the example here) to visualize my agentpy model. It's important for me to see the state of each agent, but unfortunately there are about ten different states. I've been color-coding them, following the example. However, each time the model runs, the colors are re-assigned in the order the states appeared. As a result, the color-coding is different each run. I was wondering what the easiest way to add a legend to the interactive scatterplot would be?

I've tried looking at the docs and source for the ipysimulate Scatterplot class but I was having trouble identifying exactly where the plot was being constructed, or more specifically, where it would make sense to add code for a legend.

I've also considered trying to use the ipysimulate Matplot class to rewrite the scatterplot, but it isn't clear to me exactly what the update parameter is. The source says it's a "function that takes (model, fig, ax) as input," but I don't know what the output of update should be. If using Matplot is the best route for me to take, do you have an example of what an update function might look like?

Once again, thank you so much for all your help!

[jofmi]

Hi @fl-hat,

for now, the ipysimulate package doesn't support this.

i think it is a more promising route to improve the JS scatterplot rather than using matplotlib.

you can find the source of scatterplot in https://github.com/JoelForamitti/ipysimulate/blob/main/js/lib/scatter.js and https://github.com/JoelForamitti/ipysimulate/blob/main/ipysimulate/charts.py.

contributions are very welcome, if you are interested to extend this class we can also discuss it in more detail!

[jofmi]

update: you can now write custom widgets https://ipysimulate.readthedocs.io/en/latest/guide_custom_widgets.html