to_panel issues with integers

Case 1 (i'll look for the other)

http://stackoverflow.com/questions/11329611/python-pandas-to-panel-from-dataframe-returns-weird-numbers-for-binary-variab

pm10 = pds.read_csv('pm10.csv', index_col = [0,1], parse_dates=True)
panel_exog = pm10.to_panel()['pass_ind']
