In this seminar, we will be talking about how alternative sources of data have transformed the credit scoring industry. Traditionally, customers were required to reveal details about themselves to the banks, but today, alternative sources have largely done the job of retrieving most of customer’s data. This has brought additional benefits such as improved customer experience, shortened time to loan approval and inclusion of underserved customer segments. We will also explain certain Machine Learning algorithms that have been proved successful in production. Finally, we will describe some data features that are less traditional but found useful for risk modelling.
Our speaker, Jan Sindlar, is a Senior Data Scientist at Eureka Analytics. Jan has been building risk models across Europe, India and China for the last 13 years. He has an experience in risk modelling using various sorts of data and analytical techniques to make fast and data driven decisions in lending industry. Jan's specialty is using vast array of data in credit risk scoring - behavioural, telco, mobile phone app, internet search, credit bureau, geolocation, shopping history, card & online payment transaction, p2p and many others.