This project uses the "Emotions Labeled Spotify Songs" dataset that I found on Kaggle to explore different patterns of how different songs impact people's mood. I applied various data science techniques, such as K-means clustering, correlation, Linear Regression, statistical tests, PCA, and other machine learning algorithms to identify any trends between musical and emotional features. All the code is done using Python, packages including numpy, pandas, scipy, matplotlib, and more. There are two datasets being used: one has 278k songs/rows and the other has 1200 songs/rows. Since I am not using a workstation or a computer which can handle large amounts of data, I primarily used the one with 1200 songs. The goal of this project is to both explore the dataset and also practice key data science techniques to prepare for a career in Data Science. Anything that is marked as "review" in the cell title of the code is not being used towards the project goals and more for personal practice. (See code “spotify_mood_project.py”. Report "Moodify Data Science Project" unfinished, but progress uploaded.)
mayanesen / moodify-project Goto Github PK
View Code? Open in Web Editor NEWLicense: MIT License