The Story Behind the Winning Hackathon Innovation

On February 2, 2024, the award for first place in the Housing Stabilization Hackathon, produced by Flywheel and Strategies to End Homelessness went to TenantGuard and a team led by Betsy Ehmcke. Her team included Jacob Pieniazek, Nick Ramos and Bijorn Burrell. Here’s the story behind that winning innovation,

Flywheel: Congratulations on your big win. What made you want to compete in the Housing Stabilization Hackathon?

Betsy: Our team uses data to solve problems at work every day. The hackathon seemed like a great opportunity to use the same skillset to help the Cincinnati community.

Flywheel: Tell us about you and your team. What made it click for you?

Betsy: We were a team made up of data scientists and data engineers from 84.51°. I’m lucky to work with brilliant people and was saying the whole week that I couldn’t have had a better team. These guys are so generous with their time and brought really complementary skillsets to the group. When we looked back, we couldn’t remember who had which ideas. We collaborated really seamlessly and were very excited to share our solution at the finals.

Flywheel: What were some of the first things you and did as you approach the challenge?

Betsy: With data science, it’s commonly understood that domain knowledge is just as important as analytical skill. The hackathon kicked off with speakers who have worked for years addressing this issue in Cincinnati. It was incredibly valuable and important to spend time listening to their stories. Our next move was to explore the data, taking note of its limitations.

Flywheel: Were there snags or work arounds you had to navigate? 

Betsy: A lot of people might think that it’s easy to use data if you have it – this isn’t quite true! It’s important to ask, “Is there anyone missing from this dataset?” In the case of evictions data, it misses people who are not threatened with eviction. This made it difficult to train a model that predicted yes or no, because it only had examples of ‘yes’. Instead, we trained our model on data collected by the U.S. Census Bureau from the whole population.

Flywheel: What part of your solution are you most excited about?

Betsy: I’m most excited by the opportunity to recommend specific actions depending on someone’s situation. It can be hard to ask for help and overwhelming to navigate what options are available. I’m encouraged that we’ve found a way to make that step much easier.

Flywheel: Were there any funny things that happened during the project you want to mention?  

Betsy: We joked that one of our teammates was our ‘stage mom”.  We had been working so diligently and taking this competition very seriously, so when it was time to explain our project for the preliminary-round video, he reminded us all to look at the camera and smile!