For me, the hardest part of a hackathon is coming up with a project idea. This project is meant to help those like me who just can't decide on a hackathon project to do.
What it does
By examining previous hackathon submissions on Devpost, this utility builds a predictive text model and generates a hackathon idea for you. It produces a simple HTML page (served by Python Bottle) to display the result.
How I built it
I used Python to scrape Devpost pages to generate data for the predictive model. Then, by using a simple Markov chain (not mine, source is here), I randomly generated project descriptions.
Challenges I ran into
Finding a predictive text algorithm that actually worked. I was trying to use a more sophisticated algorithm but had trouble running it.
Accomplishments that I'm proud of
I managed to get a working product finished within the 11-hour span of the hackathon without any teammates.
What I learned
I learned many Python libraries that I haven't used before. Plus, I refreshed my knowledge of some libraries that I haven't used in a while.
What's next for Hackathon Project Generator
I'd like to implement a better prediction algorithm so that the description is actually coherent.
The following unabridged Devpost text was generated based on submissions to PennApps:
look no further, because we've Got Your Back, Pack!Busy? Can't stay fit?
Inspiration: From frivolous scams like reinventing the doctor’s orders, but also difficult concepts, much cheaper to become a major issue for children is harder than other forms of apps that provides real-time visual representation of watching YouTube videos and yet despite their music being guitar, we decided to
What it does: Volunteers can be floods, earthquakes, fire, injuries, etc.) is extracted indicating the precise dosage. LocPill uses mechanical systems placed above the field. Spotted uses different superheroes (Hulk, Superman, Batman, Mr. Incredible) and Muse headband, allowing the mobile app. Through this system, the live incidents on each category. CacheTheHeat uses
How we built it: Once the system uses a MongoDB Stitch database of rlib and a virtual enclosure. The flex sensor, and image over these three points on the path smoothing with html, js, and an Arduino Uno,
Challenges we ran into: For example, "ACCOUNT_ID" might be useful, it difficult to acquire acetone, an extra team didn't have to know exactly the mobile data and tweak them with the modeling in shorts and Android Studio mobile applications of difficult. Time was developing the corresponding environments. Merging the gear in advance. This was also spent the difficulties we had one that we'd intended to the functions we were hoping for, we had dried. Hence began our features within our client's agreement. One of a crawl, we mistakenly positioned a voltage with web dashboard. Getting the level of
Accomplishments that we're proud of: The biggest accomplishment for most proud of blending together all of our design’s specs were we had even expected. This thing works, somehow. We completed and printed a backend, and a difference while hacking for those with support for us. This thing works, somehow. We were able to produce a simple camera feed. -- Achieving accurate than we would work together all decidable problems in order to leverage existing residential video
What we learned: Laser cutting, Solidworks Design, 3D printing, setting up during the Fitbit API This was surprisingly easy. We also needs to something new! We learned concepts of our first time using MongoDB Stitch, as Bluetooth Low Energy. One of signaling from the most present during the biggest lessons we'll keep in our necessity to learn something new! We were extremely valuable, our necessity to edit audio files using huge datasets like html, css, and make edits to. • Our brainwave EEG
What's next: We were