The training data for the household economic model. The dataframe is made of rows for each household cluster and columns for all independent variables (accessibility, night lights, etc.) and the dependent variable (e.g. total consumption in unit of SSD/person/day).