3 Weekends, 3 Hackathons
During this 11,000 mile journey I caught up with old friends, made new ones and discovered what excites me most about computer science — the possibility of building software to make an impact in people’s lives.
BRASATech Hackathon (Oct 31st — Nov 1st)
This hackathon was hosted in WPI (Worcester Polytechnic Institute) and lasted 24 hours. My group and I decided to address a challenge the hackathon organizers proposed: Build a tool to help professors grade tests faster. We built Testr: Teachers create tests on our platform and we generate an answer sheet for them. Teachers can use our app to grade a student’s answer sheet by simply taking a photo! The grades are stored on our platform so teachers can analyze student performance.
When the hacking ended I was really happy with what we’d accomplished. I grew even happier when we won first prize, my first ever hackathon award! I was tired, but I knew I had to recover and get ready for next weekend.
HackSC (Nov 7th — Nov 9th)
This was my first 36-hour hackathon and the first one that I worked by myself. I wanted to focus on developing a product on my own as I did with my game this summer at the MakeSchool Summer Academy. I wasn’t sure what to work on when hacking started, but I signed up to get a Myo armband. When I played with it, I realized it was limited in the programs it could control. I came up with the ideas of Gestr, a mac app that runs in the background and lets you control the operating system with Myo, switching between fullscreen apps, opening and closing apps, sending emails and much more.
It was my first myo hack. Luckily, Thalmic Labs engineer helped me to understand the myo SDK. It was my first time developing a mac app. Likewise, The Apple engineers were always available to help understanding the mac SDK. Furthermore, I learned a new language: Applescript. After 36 hours of hacking, my product was ready to show. Here is a demo video:
Hack Duke (Nov 15th — Nov 16th)
It took me 9 hours to get from LA to Durham, making Hack Duke the farthest hackathon which I traveled to attend. I had the opportunity to see people from MakeSchool whom I hadn’t seen in a while. I formed a group with 2 other brazilians and a guy from Florida State. We realized that the warning system in most colleges was neither efficient nor real-time. So we decided to solve this problem by developing an application that would analyze tweets and infer whether a dangerous situation is occurring in a certain location. We accomplished that by using a machine learning algorithm and the geolocation of the tweets. Our algorithm used a dataset of 1000 tweets manually classified as conveying danger or not. Users would use the web interface to subscribe to alert messages via SMS or email.
Alerts would be tailored to specific campuses and triggered after a certain percentage of new tweets in the area were classified as dangerous. We used Twitter, Sendgrid and Twilio APIs.
Our hack was awarded with the prize for the best use of sendgrid API. At the end of Hack Duke, I was joyful with what I had learned and accomplished.
