ActiveTrack

Machine LearningPythonScikit-learnTensorFlowPandasNumPy
github.com/haaaarsh4/ActiveTrack
ActiveTrack interface

ActiveTrack started with a practical question: can motion sensor data tell us which exercise someone is performing and how many repetitions they have completed? Using accelerometer and gyroscope recordings from five barbell exercises, I worked through the process of turning raw measurements into features a model could use. That meant cleaning and resampling data, investigating outliers, engineering features, comparing models, and tuning their parameters. I also explored signal filtering for repetition counting. The project helped me appreciate how much of machine learning happens before a model is ever trained.