It hasn't been mentioned anywhere how we have to form the dataset from the existing one in part 1 of this project
