Equivalent Functionality to `x_vars` and `y_vars` in a jointplot (splom) plotly

Hi there! I am relatively new to plotly so I apologize if this is something already covered by the docs–if so please just help me find the appropriate place.

A common functionality desired in EDA is the ability to scatter columns against each other. Plotly almost has this with the splom plot type (px.scatter_matrix(...)), however one really critical functionality that seems to be missing is the ability to filter which rows and columns get shown, a functionality present in seaborn. I had considered trying to do this manually by modifying the splom trace object, however it appears to primarily consist of a list of arrays, one for each column of the dataframe, meaning I don’t see an obvious way to remove single rows/columns as would be required to reproduce the functionality. I also looked into using subplots, and I think it might be feasible to achieve something similiar, but I then lose a lot of the convenience of px.express because I have to, for example, manually link colorbars if I am coloring each subplot by a value, and manually setup hovertemplates. Does anyone have advice on how to reproduce this functionality? I am so grateful for your help!

Hey @naveace, we had similar questions in the past, if I remember correctly without straight forward solution. You could use the forum search for these topics.

It is possible however, to hide the upper half of the matrix as well as the diagonal.