The problem description says that the last column name of the household-level training data should be “poor”. The last column for me is “country”. The individual-level training data is fine. Any idea why this is?
Have you tried selecting the column by column name (instead of just looking for the last column)?
(In pandas this can be done with df.poor or df['poor'] where df is your DataFrame.)
It is there. Column FE if you open file with Excel.
It is there in the source training data, but if does get dropped at some point in the benchmark Jupyter Notebook.