I was wondering if there are plans (or existing functionality!) to allow plotly.express plots to be created as a static image rather than an interactive plot?
The issue that I've been running into lately is when plotting large datasets, the interactive plots tend to eat up a chunk of memory. At a certain point of information (ie. too much data for one plot or a few larger plots), Jupyter tends to crash, forcing a restart of the kernel and the outputs. My workaround has been to down-sample what I am plotting, but that's not a wonderful solution, understanding that the limitation is with Jupyter + browser rather than with plotly.express. I love the visuals and quality-of-life of plotly.express over matplotlib and I would love a solution to be able to plot many charts with lots of data without the potential for a crash!
A potential type of implementation that would be interesting might follow this kind of convention:
In this case, the interactive argument would change the output of the plot to either be as it is currently (interactive=True) or as a static image (interactive=False), with a similar type of output as matplotlib.
Hello,
I was wondering if there are plans (or existing functionality!) to allow
plotly.express
plots to be created as a static image rather than an interactive plot?The issue that I've been running into lately is when plotting large datasets, the interactive plots tend to eat up a chunk of memory. At a certain point of information (ie. too much data for one plot or a few larger plots), Jupyter tends to crash, forcing a restart of the kernel and the outputs. My workaround has been to down-sample what I am plotting, but that's not a wonderful solution, understanding that the limitation is with Jupyter + browser rather than with
plotly.express
. I love the visuals and quality-of-life ofplotly.express
overmatplotlib
and I would love a solution to be able to plot many charts with lots of data without the potential for a crash!A potential type of implementation that would be interesting might follow this kind of convention:
In this case, the interactive argument would change the output of the plot to either be as it is currently (
interactive=True
) or as a static image (interactive=False
), with a similar type of output asmatplotlib
.Loving the work on this library!