GraPics Determines Your Friendship Base On Photos
We asked what happened at a HackNY Hackathon we participated in.
HackNY is a New York City-based initiative seeking to create and empower a community of student-technologists. Created in 2010, the organization is organised by faculty at both New York University and Columbia University and aims to federate the next generation of hackers for the New York innovation community. HackNY holds semiannual, 24-hour “hackathons,” and runs a summer Fellows Program which pairs students in the computational sciences with local startup companies.
Each semester, hackNY holds a 24-hour coding marathon. The events are open to all students, not just of Columbia and NYU. The hackathons begin with brief presentations from startups that make their APIs available to the participants. The students are encouraged, though not required, to work together to come up with creative solutions or build new features before the time runs out. The hackathons conclude with a presentation of the students’ work, and judging.
1. What prompted you guys to come up with GraPics and what’s the inspiration behind it?
Our idea made use of graph structures to visualise tags from our pictures, so naturally graph + pics = GraPics! Our tagline is that a picture paints a thousand connections (yeaps, pretty deep).
2. What are the key features of GraPics you would like to highlight?
TL;DR Connecting people by their experiences.
Extracting all your photos from your Google Drive and with machine learning, tagging each of those photos to keywords that best describe the photo. We then visualise those keywords onto a graph network and see how similar you are to your friends/strangers solely on the pictures they have on their Google Drive.
3. How did you guys become a team/meet? How are your team dynamics/roles?
Mark and Brenda (they are housemates) signed up for the hackathon together and were intending to find other teammates at the hackathon. We later found out that Kenneth joined the hackathon as well but didn’t have a team, so it was really a match made in heaven.
Mark was responsible for setting up the Google Drive API and ingesting the JSON to be sent to the machine learning model. Kenneth built that machine learning model and Brenda was in-charge of visualising the outputs of the model.
4. What are the challenges faced during the hackathon?
We all had a pitch session with Josh Futterman earlier in the day so we hadn’t gotten much rest the night before and at the hackathon itself. We also didn’t settle on our idea or stack till 6 hours into the hackathon which didn’t give us much time to actually build out the project fully. We also made the naive mistake of focussing the entire idea based on two or three APIs even though we hadn’t had experience with those APIs. That meant that we couldn’t use alternative APIs if we were stuck.
5. Did anything interesting/funny happen while working on the project?
We installed more tools than we used (I mean like, who doesn’t like to try out new IDEs?!). We played a game of werewolf at the stroke of midnight with the other hackers, which was pretty cool. Someone lost his/her’s wallet towards the end of the hackathon, but luckily he/she got it back by a stroke of sheer luck.
6. Any tips for aspiring entrepreneurs or students going in to NOC?
Mark: You are no longer a student when in NOC, so take charge of your own personal development and be willing to step out of your comfort zone.
Kenneth: You gain some, and you lose some (sleep)
Brenda: Don’t be afraid, go out there and try everything (including the food, free food just take). As for hackathons, try not to come up with an overly complicated idea. Try to come up with something that you’d be able to complete within the limited timeframe.