Counting Leading NaN Values in Original Columns and Non-NaN Values in Extra Columns with Pandas DataFrames