`glum` support for `maketables`

[s3alfisc]

Hi @ quantco team / glum devs,

I have set up a draft PR to add glum support for maketables, which we hope can become in Python what modelsummary is in R by providing a general interface to produce regression tables across different packages / model classes.

If one of you has 10 minutes to spare, would you be so kind to take a look at the PR? py-econometrics/maketables#13

Mostly, I am wondering about two things:

• Is it possible to somehow obtain the name of the dependent variable when the scikit-learn syntax is used (as opposed to the formula API)? I would assume no?
• Is it possible to retrieve information on whether the variance covariance matrix is clustered, and if yes, by which variable? If no, would this be something you'd consider adding?

@dsliwka FYI

[MarcAntoineSchmidtQC]

Thanks @s3alfisc. This will be very useful for the community. I'll take a look at the PR.

Regarding your questions:

1. Currently, we could store this information if y is passed as a Series with a name or a DataFrame with a single column. It would require adding an internal parameter.
2. Similarly, we could also add an internal parameter. My draft proposal would be to add the name of the variable, if available, or True, if standard errors are clustered but we don't know the name.

[s3alfisc]

This sounds like a good approach to me! =)