Here is a little Gist I did for exploratory data analysis

[dartdog]

Not really a package, too simple, but I think useful, I use pandas-profiling 1st to check for an overview of data quality and covariance then try this which shows how separated your variables are between classes I'd like to rank the variables by that factor(variance between classes) and include counts in a master dataframe, but have not got that done yet. IMHO that provides a good guide for feature selection and a rough guide for feature evaluation for Machine learning? 

Thoughts?

https://gist.github.com/dartdog/67cd600c62fb7d3ef118af3ca3f02f7a