What is your project about?

The financial services industry has always been heavily data driven. Ever since its beginning, the delivery of its core services has relied on data analysis for modeling customers’ credit worthiness, predicting payment defaults, computing insurance premiums...etc.

Traditionally, financial models primarily used data variables which were exclusively available to banks, such as a person’s past financial transactions. In my project, I study how we can utilize user-generated digital footprint data, as an alternative to traditional financial data, in order to deliver better financial services. In this context, we define digital footprint as the trail of data each one of us inadvertently generates whenever we are online. These can be related to our online shopping habits, social media interactions, the devices we use, telecommunication usage patterns and much more.

In this project, I take an applied approach where I design, implement and deploy various machine learning algorithms in order to model the financial services I study, and accurately predict their outcome.

Why did you want to work with this specific topic?

This project was a natural extension of my professional and academic background: on the one hand, I have built a solid foundation in finance, statistics, and business development throughout my studies. On the other hand, I got the opportunity to work within the technology industry early in my career, which made evident for me the impact digitalization and data-drive technologies are going to have on all aspects of society and business. Therefore, as I was exploring various research topics, I particularly focused on applied topics which lie at the intersection on financial technology and machine learning, as is the case for my current project.

What would you say is the goal of your project?

The primary goal of this project is to study the technical and business enablers for using user-generated digital footprint as an alternative to traditional financial data in the delivery of financial services. To achieve this, I build machine learning models that can accurately predict the outcome of various financial services such as the customer’s credit score and the electricity spot prices. I also study the underlying economic and market pre-requisites which can motivate the real-life adoption and deployment of digital footprint-based models.

Do you have a target audience?

The results of this project can be used by the providers of financial services. These include traditional incumbent banks which want to become more competitive and agile, incumbent technology companies which aim to enter the financial services market, as well as the emerging FinTech startups which want to build their businesses around these alternative digital-footprint models.

Furthermore, this project targets policy makers who need to draft new fiscal policies that regulate the proper use of digital footprint data in the financial services industry. Finally, it is very important for us to also reach individual citizens, in order to raise their awareness about how their own personal data can potentially be used against them when applying for services such as loans and insurance policies.

Have you, or do you think you will, experience any challenges during your work?

At the beginning of this project, it was important that I acquire a solid understanding of the technical advances within machine learning, as well as the various economic and market dimensions of the financial technology industry. While this surely required a steep learning curve, it was also a very rewarding one. In the remainder of this project, where my work focuses more on data modeling and algorithm implementation, I need to ensure that I have timely access to the various datasets required for training the models and testing their accuracy.