Creation of a denormalized table that contains behavioral indicators on the customers, calculated on the basis of transactions and products ownership. The aim is to create the features for a possible supervised machine learning model.
In particular the features to create, for each client_id, are the following:
- Age;
- Number of outgoing transactions on all accounts;
- Number of incoming transactions on all accounts;
- Outgoing amount transacted on all accounts;
- Amount transacted on all accounts;
- Total number of accounts held;
- Number of accounts held by type (one indicator per type);
- Number of outgoing transactions by type (one indicator per type);
- Number of incoming transactions by type (one indicator per type);
- Outgoing amount transacted by account type (one indicator per type);
- Incoming amount transacted by account type (one indicator per type).