One time a friend of mine had to organize a callout for Purdue Hackers, and he was expecting 40 people to come since about 100 people clicked going on Facebook, but almost everyone attended the event and he had to order more pizza.
What it does
It predicts approximately how many people will be visiting your event according to the previous history.
How I built it
Machine learning to generate data-regression and make approximations based on the history. And used python+JS to generate pretty plots on web.
Challenges I ran into
I had no data to predict anything. I had to run around during the hackathon and count how many people attended for each tech talks and used that data to generate relatively more reliable data set.
Accomplishments that I'm proud of
I did it buy myself without teammates. I learned front-end developing for the first time at the hackathon and it was awesome.
What I learned
A LOT OF JS