Back in September 2017, I was really tired and learning about Maximum Likelihood and yearned to do some more machine learning. For the longest time I wanted to scrape my gym’s twitter account to get information about when the gym was busiest.
It turns out that doing so is incredibly easy, and so born was GyMBo , my gym monitoring bot.
GyMBo uses some very naive heuristics to parse the gym’s tweets, and then uses polynomial regression to fit a quadratic function to the data. Simple? Yes. Effective, also yes. GymBMo has a mean absolute error between 11 and 13 on any given day. That is pretty good for something very simple.
I’ll be posting GyMBo on GitHub soon, that way you can point the script at your own gym’s twitter account and have your own personal GyMBo. I plan to eventually use some more flexible models as I collect more data.
Did I mention GyMBo tweets? Here is where you can find GyMBo on twitter. He likes to tweet once an hour.