ds-simple-portfolio
Project 1: K-Pop Analysis
Overview
-
Created a web application that returns the predicted number of hours one listens to K-pop on a daily basis using FLASK (MAE ~ 1.2 hours).
- Engineered features from the text of each column.
-
Explored the data to analyze the relationships among the features (or variables).
- Built five different regression models - linear, lasso, ridge, random forest, and XGBoost.
- Optimized the random forest and the XGBoost model using GridsearchCV to find the optimal parameters.