Conclusion

Fitbit performed the exact analysis you just completed in this project. They used 100+ billions of hours of data from their user population in 2018. Read more about it here.

As wearables become more widespread in medical research, these kinds of analyses can play a large part in the discovery process. With large enough datasets and more informative clinical metadata (in this case, we only used age group), wearables can be used to discover previously unknown phenomena that may advance clinical science.

Before doing any analysis, however, engineers and data scientists will have to process the large amounts of raw data that wearables collect. By completing this course, you have taken a first step in this direction :)

Food for thought:

Does the above graph look similar to any graph created in this project?
Remember, the dataset we used had only 36000 hours instead of 100+ billions of hours, how would this affect our results if we had 100+ billion hours of data?