InVenture’s chief data scientist, Peter Sugimura, spent his early career building models of the Arctic Ocean to study climate change. Now he’s using data to unlock financial access in East Africa and beyond.
Learn more about what motivates him — and the other Peter who inspires him — in our Q&A.
Q: Much of your previous work as a data scientist was around climate change and clean energy. What made you interested in financial access?
A: I’m always looking for roles where the science is interesting but where there’s also a strong social mission. My most recent roles were with solar energy startups — I was very driven by the challenge of climate change and how science could help. Financial access was probably the last thing on my mind. But when I learned about InVenture’s product, I was immediately drawn to the impact it could have on people’s lives. I also saw that the data science side of the work was enormously interesting — and that the potential for data to really innovate in these markets was huge.
Q: What is most exciting to you about the data you’re working with at InVenture?
A: In emerging markets, lack of data is one of the major barriers to financial access. There is little to no financial history for the majority of people in these economies, meaning they can’t prove their creditworthiness or obtain traditional financial services. What InVenture is trying to do for this population is build a credit history where previously there was none, using the data that’s seamlessly collected on their smartphones. We’ve learned that we can build an accurate measure of a person’s ability and willingness to repay a loan based off of factors that you may not immediately think of as financial indicators, including communication and social data. In many cases, we’re able to prove that a person is creditworthy who might otherwise be overlooked.
What’s fascinating about mobile data — and the reason it works so well for these purposes — is that our phones are windows into our identities and behavior and one of the strongest links to who we are. This is an area of data science that hasn’t really been tackled before, and we’re using it to benefit people in a very direct, very tangible way.
“Meeting our borrowers in person and understanding the real impact of credit on their daily life brought new meaning to the work that we do. It also gave us a window into the human side of our data.”
Q: On a trip to Kenya a few weeks ago, you met customers who have benefited from InVenture’s Android app, Mkopo Rahisi. What was that experience like?
A: It was my first time in Kenya and so it was fascinating to get a better understanding of the markets we’re working in. We hear from our customers regularly and share their stories through our Wall of Love, but meeting our borrowers in person and understanding the real impact of credit on their daily life brought new meaning to the work that we do. It also gave us a window into the human side of our data and the impact of our credit decisions. The models we build are, by their very nature, discriminatory: the goal of the model is to separate good borrowers from bad. But seeing how people use the loans, and understanding the consequences for their lives if they aren’t able to receive loans, gave me more incentive to build as good a model as possible — one that minimizes false negatives and proves people’s potential.
One borrower I met, also named Peter, made a huge impression on me. He was the classic hustler, juggling two full-time jobs as a taxi driver and barber, good days from one job balancing bad nights at the other, borrowing from one to pay the bills in the other, barely breaking even. A few years back, he had owned a barber shop, but an emergency forced him to give up the shop, and now he was building his way back to stability. His story stood out because it showed how tenuous his grasp on a stable life was, and how Mkopo Rahisi was helping him accelerate his goals and ease the day-to-day stress of hustling.
Q: Mobile data collection is not without controversy. As a data scientist, how do you see the issue?
A: I think every person has a different level of comfort with data sharing, myself included. It’s important that our customers understand what data is being collected and how it’s being used so that they can make the decision that’s right for them. When our customers install our Android app and begin the loan application process, they have to elect to allow access — we do not assume access is granted. At this point, the customers understand that we’re using their data to make a credit decision and can properly evaluate the cost-benefit to them. It’s important to remember that the data is a means to help our customers access credit rather than an end in itself.
The other point to emphasize is that the continual flow of data helps us build a dynamic model that adapts to individual circumstances and ensures we aren’t building a one-size-fits-all solution. Because we can tailor our product to each customer, we can avoid pitfalls that might trap borrowers in debt or create an otherwise unsuccessful borrowing experience. It’s a win for the customer and a win for us, and it’s only possible because of the richness and diversity of mobile data.
“One of the most enjoyable parts of my job is uncovering new data nuggets that help us improve the model and reach more people. I’d love to go back to customers we initially declined and be able to offer them assistance.”
Q: What’s next for data at InVenture?
A: In the near term, I hope we can continue to grow our product and help more people access credit. One of the most enjoyable parts of my job is uncovering new data nuggets that help us improve the model and reach more people. I’d love to go back to customers we initially declined and be able to offer them assistance.
In the long term, beyond credit, there are so many possibilities. The type of data we’re collecting will help us open access to other products and services that can really change people’s lives. The opportunity is tremendous and I look forward to whatever’s next.