My teammates and I were trying to decide on a project and we knew what we might want to use but not necessarily had an idea. We figured we'd take that and try to do something with it
What it does
You list out the technologies you plan on using for the hackathon and it will return to your chances of winning on a scale of 0 to 100
How I built it
I used scikit-learn's SVM module and trained a model on all of devpost's previous projects in order to try to find relations between the winning the projects and what they used.
Challenges I ran into
This was my first time doing anything with machine learning so trying to figure out what to use and how to do it was the biggest challenge for me.
Accomplishments that I'm proud of
I'm glad that I've finally started to understand SVM's and neural networks a bit more, even though I didn't implement a NN for this project.
What I learned
Learned a lot about SVM's and neural networks just from the research I was doing in order to solve the problem.
What's next for CrystalBall
Potentially find more statistical analyses of the data set in order to find more accurate trends and maybe even use recurrent neural networks to run through all of the project descriptions as well to allow the bot to literally create ideas rather than giving your chance of winning.