[Python3] Predict type of people base on their expenses

The subject uses bank-like data. The goal is to predict the categories of people based on their expenses. You must work in Python3 and deliver a clear and commented python file that execute the process. Librairies to use : scikit-learn, pandas. Deadline : 31th May. You have 2 files : 1. The first one has 2 columns : "Name of operation"; "Category of operation". The goal is to train a model to be able to categorize the expenses (in categories such as "credit card", "incoming transfer") base on their names. This involves a bit of text-mining to clean and work with the rough operations (stemming...). 2. The second one has 4 columns : "ID of customer"; "Category of customer"; "Name of operation"; "Amount of operations". Here you first must to create a column "Category of operation" by using the process you create in step 1. Then, aggregate the data by category (min, max, mean, whatever might be useful) to be able to predict the column "Category of customer" for each "ID customer". Details : Kaggle experience in banking problems is a plus. The goal is to obtain something better then an already-existing code. PLEASE describe briefly the similar projects you worked on to boost your chances of selection !