Hi, you need to find the features in the other data files. So in e.g. transactions there are transactions for the IDs you'll find in test. Likewise in customer you'll find customer info for the IDs in test.

I think you need to extract features from the other files and align them all on ID, model and predict that, and then use the IDs from test only to report or print your submission file based on the features found elsewhere.

The values you find in the Train.csv are the income targets. So you have to predict the incomes for the customers displayed in the Test.csv. (Only the customer id info is useful from that data.)