99P Labs 2022 Data I/O Recap

Introduction

Data I/O is central Ohio’s premier one-day datathon. The Big Data and Analytics Association, a student organization at The Ohio State University, hosts it with the support of OHI/O. This year it took place on November 5th at Curl Viewpoint on The Ohio State University’s campus. Curl Viewpoint is a beautiful space perfect for the event’s size. The event ran from 10 am to 6 pm. As such, it was a single-day hackathon where competitors got about six hours to work on their projects.

In addition to sponsoring Data I/O, 99P Labs team members also participated as mentors and judges. The purpose of this blog post is to share our experience from the event.

Approximately 50 students participated in 12 teams. During the event, each participant worked on a challenge with their team and then decided if they had made enough progress to present their work. Getting over the initial fear of presenting is difficult since no project is ever perfect, and with such a limited amount of time to work on it, imposter syndrome can take over. That’s why we were delighted that 11 projects signed up to present. In most hackathons, there’s a significant drop-off between those who start and those who finish. Seeing so many teams demonstrate their resourcefulness and grit while finishing their projects was heartwarming.

The Problem Statement

Sponsoring a hackathon includes the opportunity to offer a challenge for participants to complete. 99P Labs challenge was called the Mobility Pattern Analysis Challenge.

Students were given a sample of our V2X data, which is vehicle trip data, as the data for our challenge. The V2X data is collected as part of the US 33 Smart Corridor initiative. The data is a feed from V2X instrumented vehicles and captures various external interactions. If you’d like to read more about the V2X dataset, please check out the 99P Labs Developer Portal.

For our challenge problem statement, we asked the competitors to identify different groups in the data. Then, we told them we would examine the characteristics, methods, and reasoning they used to create the groups. By group, we meant any differentiating factor. Some examples of differentiating factors are (a) rural/urban, (b) speed, (c) trip duration, (d) trip destination, (e) population density, (f) vehicle density, and (h) time of day. In addition, we explained how these examples were not exhaustive, and we hoped they could provide some more examples.

The challenge problem statement ended by asking the challengers to answer the following questions in their submissions:

1. What insights were you able to gain from your analysis?
2. What criteria did you use for differentiating groups?
3. How did you develop those criteria?

The Event

There was a wide range of skill levels represented in this event, making it great for students of all levels. Among the participants, some were entirely new to Python and data, while others were highly experienced. Despite this difference in skill sets, the atmosphere remained friendly. The event fostered a very inclusive atmosphere where people who were just learning about data had the opportunity to network with other people who shared their interests.

A great thing about this event was we could answer many more questions because there were about 50 students, unlike our previous hackathon, which had approximately a thousand attendees. This more relaxed atmosphere allowed us to field more questions and interact with the participants in a much more personal way.

The Submissions

As I mentioned earlier, we were very pleased that most students who participated in the challenge submitted a project. There is nothing more disheartening to see than a team that spent most of the day working on a project only to be overcome by anxiety that their project isn’t good enough and decide not to submit it in the end. We tried to explain the immense value of the experience of turning in the project and presenting, even if your project wasn’t finished or as good as you’d hoped, and we were happy when all the teams on the fence decided to submit and present.

Despite only having a short amount of time to work, the teams came up with some very impressive insights and visuals.

Just to go over a few examples of the insights, a few looked at the relation of speed to windshield wiper levels. Specifically trying to determine why speeds were generally lower when the first level of wipers was on compared to a higher level of wipers being on. One group built a dashboard in Python Dash that allowed you to select specific columns from the data set and check correlations between them. And finally, one group looked at speed in relation to the day of the week and was trying to hypothesize as to why there was higher speeding on Sundays compared to Fridays. Their hypothesis was it had to do with the increased level of traffic on Fridays compared to Sundays.

While all the teams that turned in and presented did a great job, the event still had to determine winners. A team of 99P Labs and OSU faculty determined the four winning teams. Below are pictures of the four winning teams: (top left to right) 1st Place: Paranormal Distribution, 2nd Place: Pomerene Pools, (bottom left to right) Best Visual: akhilsanjay, and Best Insight: Dataholics.

Conclusion

In conclusion, I hope it was clear that we had a blast participating in this event. All the students working feverishly to solve problems created a contagious inspiring energy in the room. It’s great to be able to participate in events like this. Taking part in such events helps us work towards our goal of building a developer community that can address future mobility problems. Check out our blog post for more information about our developer community. If you’d like to reach out, you could connect with us on LinkedIn or Medium. You can also reach us via email at research@99plabs.com. Thank you so much for reading, and please do not hesitate to reach out with any questions or concerns or if you’d like to discuss collaborating with 99P Labs.