Coding A Fitness Tracker

Role

I was the lead of developing the python scripts to analyze sensor data

Project goal

Create Python scripts to process, visualize, and model accelerometer and gyroscope data to create a machine learning model that can classify barbell exercises and count repetitions.

Solution

1- Read all the separate raw CSV files, process them, and merge them into a single DataFrame.

2- Create data visualizations to better understand the accelerometer and gyroscope data for the different exercises.

3- Check whether there are any outliers (extreme values) in our data that we want to remove using various methods.

4- Filter subtle noise (not outliers) and identify parts of the data that explain most of the variance. Then add numerical, temporal, frequency, and cluster features.

5- Experiment with feature selection, model selection and hyperparameter tuning with grid search to find the combination that results in the highest classification accuracy.

6- Count repetitions and evaluate it against the dataset.