Using EDA Python packages

Sometimes it's helpful to create some specific EDA plots and statistics to investigate features of interest, but often, it's helpful to run an auto-EDA package on our data as one of our first steps. There are a host of different EDA packages in Python (and R), but we'll stick to just covering pandas-profiling. This is a convenient package that creates an EDA summary with only a few lines of code from a pandas DataFrame. Once we have our data loaded, we load the ProfileReport function from pandas-profiling:

from pandas_profiling import ProfileReport

Since dashes are not allowed in module names, we need to use an underscore for the library name, pandas_profiling. Once we have this loaded, we can create our report and display it:

report = ProfileReport(df)

Within Jupyter Notebook, we have a few options for display. We can simply print out the variable in a Jupyter Notebook cell like so:

report

Or, we can use report.to_widgets...