Abstract
Microfinance has known a large increase in popularity, yet the scoring of such credit still remains a difficult challenge. Credit scoring traditionally uses socio-demographic and credit data, which we complement in an innovative manner with data from Facebook. A distinction is made between the relationships that the available data imply: (1) LALs are persons who resemble one another in some manner, (2) friends have a clearly articulated friendship relationship on Facebook, and (3) BFFs are friends that interact with one another. Our analyses show two interesting conclusions for this emerging application: the BFFs have a higher predictive value then the person’s friends and secondly, the interest-based data that define LALs, yield better results than the social network data. Moreover, the model built on interest data is not significantly worse than the model that uses all available data, hence demonstrating the potential of Facebook data in a microfinance setting.
Acknowledgements
We would like to thank the data science team at LenddoEFL for providing us with the data and valuable feedback.
Notes
No potential conflict of interest was reported by the authors.