Additional Namespaces about pyjanitor

[ericmjl]

Specifically for the “finance” sub module, I’m comfortable building it up slowly, one function at a time. That’s how good software gets built, after all 😄.

[ericmjl]

As for the summary sub module, let’s consider first whether it’s better to just wrap pandas_summary (https://github.com/mouradmourafiq/pandas-summary/blob/master/README.md).

On the other hand, summarizing a dataframe doesn’t return a modified version of the original, but a new dataframe that describes the original. Pyjanitor so far hasn’t done much with creating new DataFrames, with the exception of getting x/y columns for ML purposes. We can add it, but yes, it should be a separate namespace.

[szuckerman]

pandas_summary looks interesting, but I think that module looks more like it's looking at data integrity (i.e. null-values, uniques, min, max, etc).

Along those lines, I think that integrating PandasSchema, or something like it, would be really helpful.