fitbit
During my master Data Science and Entrepreneurship I was looking at various ways of leveraging data for the purpose of entrepreneurship. I realized that my FitBit wasn't all that accurate and at times displayed weird values. For that reason I decided to create a model that could predict what my heart rate would have been if it were missing (e.g. due to a bad signal).
The data itself is not included. However, the notebooks show how I went about doing my analysis and give a nice overview of what I did. Notebook "3. The Final Product" gives a nice summary of preprocessing the data, doing EDA, modeling and interpreting the results for a business case.
This was the very first project I did during the pre-master. Looking back, there are some things I could have done differently. For example, I did a simple 10-fold CV to test my models which might lead to overfitting as the folds are chosen randomly.
