One day I was watching a Formula 1 race and I thought if I can predict which track is just by telemetry, ignoring car speed, track length and lap time. And that's what this project is about.
The data was collected from the Ergast Developer API and Fast F1 that is a Python Lib. With this lib we can get some telemetry from the car, but only trivial ones like speed, rpm, throttle, lap time, etc. But I used only the percentage of throttle and the gear. But even with that, I still have a problem. The data is a time series for each feature and I want to eliminate that to remove the time variable. So I just used the mean of each feature for each sector of the track. So I have 3 sectors for each track and 2 features(throttle percentage and gear) for each sector. So I have 6 features for each track. On get_info.ipynb you can see the process to extract the data from the lib and make it to a csv file.
Before go straight to the models and see the results, is good to make an Exploratory Data Analysis and I made a simple one on EDA.ipynb.
I used Decision Tree and SVM to predict the tracks. I used the Decision Tree because it's a simple model and it's easy to interpret. And I used SVM because it's a powerful model and it's good to compare with the Decision Tree. I used the GridSearchCV to find the best parameters for each model. The results are on formula.ipynb. I didn't use Neural Network because I found a particularity on the data that I understand there is no need to use.
Hope you enjoy this!