Analyzing Taylor Swift Song Data

• 

  Visualized KMeans Clusters

• 

  Embedded Taylor Swift Songs by Album

Inspiration

I was inspired to analyze Taylor Swift data for my project because I am a huge Taylor Swift fan and I thought this would be a great opportunity for me to see her songs through a new lens.

What it does

This project accomplishes three things:

• Figures out the top 3 albums based on popularity statistics for each individual song
• Finds the "average" Taylor Swift song and visualizes similarities between songs by album
• Determines which musical characteristic is the best for predicting what album a song is from

How I built it

• Built using data programming skills learned in CSE 163 this quarter
• Used Python for the coding, a CSV file for the data, and various libraries for visualization and data processing

Challenges I ran into

• Figuring out how to visualize multidimensional data easily
• Coming up with a good plan for solving the machine learning question with good accuracy

Accomplishments that I'm proud of

• Using the TSNE library to compress and visualize multidimensional data
• Ensuring that both the training and testing set contain songs from each album and using a RandomForestClassifier to improve prediction accuracy

What I learned

• How to apply skills to a real world project
• How to overcome unexpected challenges and research solutions
• How to write about and present coding solutions

What's next for Analyzing Taylor Swift Song Data

• Further questions at a similar level (most unique song, least popular album, etc.)
• More complex machine learning techniques to improve accuracy further (hyperparameter tuning, model depth, etc.)
• Comparison of Taylor Swift's discography with another artist's work

Built With

• python