Relationship managers for small and medium enterprises (SMEs) have a large client portfolio. This makes it difficult to keep track of their clients. As a result, banks serve SMEs in a reactive manner. Banks want to help their relationship managers to better serve their clients. A common client request is a (renewal on their) business loan. As part of the MVP, this project uses machine learning to suggest to the relationship managers which customers are likely to apply for (renewal of) a credit, and when they are likely to do so. The goal is a solution to help relationship managers to serve their customers more proactively. Hence, the project aims to develop a model that predicts which clients will be applying for a loan

The data covers a time period of 32 months in total (January 2014 until August 2016). The first dataset “customers” contains a subset of customer data like credit volumes, debit volumes, number of transactions, etcetera. The second dataset “credit_applications” contains information regarding historic credit applications. The “credit_application” field indicates that in the given month (see column “yearmonth”) has a value of 1 if a client in that month applied for credit with ABN AMRO and otherwise it has the value 0. Field “nr_credit_applications” indicated how many times a client applied for credit in a given month.