Amit Kumar Mishra

Here, I analyze performance data generated from Oura ring. The process involves connecting to the database to pull the raw data, cleaning it and then analyzing it using machine learning models.

Here’s the connection process:

Data munging and preparation phase

There are tables with more than 50 KPIs. Based on what the team required, I’ve performed some data prep using joins, removing the null values before the modelling phase.

Model fitting phase using Ridge Regression to predict Readiness Score

Now that we’ve created the model using training and test data, let’s put it into something more usable. We can do that by deploying it on Streamlit with a simple interface for taking inputs to generate the readiness score.

Deploying the model on Streamlit

We get a predicted Readiness Score using three KPIs — Average Resting Heart Rate, Sleep Score, and Activity Score.