Hi,
I have observed that running go.Box(...) in Jupyter Notebooks results in storing all the raw data (x, y) in “application/vnd.plotly.v1+json” JSON fields in .ipynb files. I suppose this means that the actual aggregation only happens in JavaScript and not Python. This is however terribly wasteful for large datasets (huge notebooks which have to be saved or kept loaded in the browser). Ideally, only the necessary values for generating the plots and not the raw data should be stored.
Is there any general way how to do this? Or at least make box plots from statistics and not raw data?
Thanks!