/Analysis-of-Bank-Customers-Using-SQL

Analysis of a bank dataset to extract behavioural features to be used in a Machine Learning model

Analysis of Bank Customers Using SQL

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).