Possibilities of Using Digital Traces to Assess the Credit Potential of the Russian Population

This research discusses such a concept as a digital footprint and the effectiveness of its use to improve scoring models. The database of Bank X is considered as a database, and the influence of the digital footprint is projected on the predictive ability of the scoring model for assessing the credit potential of customers in the online segment. The work is developing a model to counteract potential fraudsters, in which the digital footprint data shows its effectiveness. In addition, a "classic" scoring model for customer applications is being built, in which the introduction of digital footprint data significantly increases the efficiency of assessing the bank's customers. At the last stage, a model is built for payroll clients, as well as for clients who have applied for a loan for the first time. The fact is that people who came for a loan for the first time are difficult for the bank to assess and the introduction of digital footprint data improves the quality of assessment by more than 25%. As a result of the work, the results of the completed tasks and tested hypotheses are summed up, and the practice of real use of digital footprint data in banking practice in 2021 is discussed.