Oura Ring vs. Fitbit Charge 4 Project Type: Applied Data Science / Machine LearningFocus Areas: Wearables analytics, feature importance, model interpretabilityTools & Methods: Python, Linear Regression, Permutation Feature Importance, SHAP Project Overview Consumer wearables promise actionable insights into sleep quality and daily readiness, yet their scoring algorithms remain proprietary and opaque. This project reverse-engineers and
Gym Visit Analytics — 2024 Attendance InsightsIn this self-directed project, I leveraged attendance data from my local gym to quantify my workout consistency and membership cost efficiency for 2024. I collected visit logs from the gym’s web portal and asked key questions: Was I consistent in making it to the gym? How many visits did
This visual is the output of work to identify factors that may impact the age of marriage. The original hypothesis of this data was that cost of living played an important role in when people choose to get married. However, after finding minimal correlations between the cost of living and population to the age at
https://www.kaggle.com/pmrich/obesitydataset-eda-data-prep-ml-hypertuning This data comes from the UCI Machine Learning Repository. This dataset include data for the estimation of obesity levels in individuals from the countries of Mexico, Peru and Colombia, based on their eating habits and physical condition. This notebook explores several popular machine learning classification models to predict the weight classification of patients. This notebook provides a walk-thru of
This is a Jupyter notebook I created and published on Kaggle to demonstrate how clustering can be used for customer segmentation. https://www.kaggle.com/pmrich/clustering-approaches-k-mean-birch-agg Using clustering and credit card data, I can group the credit card users into three (or more) distinct groups that could be used for various business applications such as marketing and promotional campaigns
