Tries to shrink your Pandas column dtypes with no data loss so you have more spare RAM - View it on GitHub