Numina WiTNY Winterns Gamify City Streets

This winter, Numina welcomed five bright interns (Cristell, Elena, Iris, Jasmine, and Sherien!) for a three-week internship as part of the Women in Technology and Entrepreneurship in New York (WiTNY) program. We were excited to work with these CUNY students on a project that challenged their technical and creative abilities. Their final project exceeded all our expectations, combining lean prototyping, social media, game design, and data analytics to create a fun and useful tool to address one of our company’s most pressing needs.

At the start of the internship, we posed a question to our winterns: How might we engage and incentivize local communities to contribute training data to Numina? Numina uses computer vision-based sensors to collect anonymous movement data from objects such as pedestrians, bicyclists, and vehicles to understand traffic and safety patterns. To do so, we first trained our computer vision algorithms on large, publicly available image datasets that include images of pedestrians, cars, and so on. However, in order to detect other important objects in cities, such as trash, emergency vehicles, or abnormal crowds, we need to collect more images of them to “train” our algorithms.

Over days of research and ideation, our winterns tested different technologies and approaches to this question using QR Codes, Mailchimp, and other web tools. Ultimately, they decided to design Play My Streets, a game that challenges city dwellers to submit photos of objects like trash bags.

They ran an experiment comparing game engagement with campaigns on different platforms: a Web landing page, Google Forms, and Twitter. After observing the highest engagement on Twitter, they launched an official game on Twitter and created a funny promotional video to explain the game.

Finally, using Twint, a Python-based Twitter scraper, the winterns were able to continuously monitor uploads to the #PlayMyStreets hashtag, maintaining a leaderboard of Twitter users and archiving images for our team to use for our trash detection algorithm.

In just a few weeks, our winterns completed an end-to-end project including research, design, user testing, and prototyping! They documented their progress in a presentation, which highlights how methodical and comprehensive their approach was.

Wintern-Deliverable-Presentation

We were impressed and inspired by the work of our winterns. The new ideas and positive energy they each brought into the office every day was also a treat!. We look forward to using Play My Streets to collect more city training data and participating in future WiTNY programs.