Mainstream Artist Predictor

Inspiration

We found inspiration from the fact that we are all people obsessed with music, and wanted to see how some of our favorite artists — and even ones we haven't heard of — might grow in the next couple of years.

We all have extensive backgrounds in the realms of Python, Java, and CSS, and we wanted to use the techniques we've learned to predict how these artists might evolve over time.

What it does

The project essentially allows users to select which artist they want to look at on the website, and see what their predictive score will be for the future. This allows them to see graphs of the artist's current popularity score, and how they are expected to grow in the future with their current numbers and datasets.

How we built it

We built the backend of our project using Spotify's API via the Spotipy Python library. We extracted data on the top 50 most popular artists, including:

Artist name
Total stream count
Popularity score
Monthly listener count

We then used tools like Python, Pandas, NumPy, and Matplotlib to construct multiple dataframes and visualizations that showcased each artist’s career trajectory, how their popularity compares to their follower count, and other key insights.

For the frontend, we used:

JavaScript
React
HTML
CSS

This allowed us to build a dynamic webapp that highlights artist data and includes interactive graphs showing projected popularity trends.

Challenges we ran into

Some artist data was incomplete — particularly around career start dates. This required manual lookups and data cleaning to ensure accurate modeling.
We had to make decisions about modeling techniques: Should we use linear or logarithmic regression? We realized some artists grow steadily, while others spike rapidly — and our models needed to reflect that diversity.
Designing the frontend-backend interface posed some challenges, especially when integrating Python-generated plots with a JavaScript frontend.

Accomplishments that we're proud of

Successfully pulled real-time data from Spotify’s API
Created multiple insights using data science techniques
Built a functional and visually appealing frontend with React
Developed predictive visualizations using real-world artist metrics
Combined our love for music with technical skills to build something meaningful ## What we learned
How to work with third-party APIs and handle missing/incomplete data
How to use libraries like matplotlib, seaborn, and pandas to turn raw data into insights
How do different regression models (linear vs. logarithmic) reflect artist growth differently
The importance of collaboration between frontend and backend development
How to connect a technical project with a passion for music ## What's next for Mainstream Artist Predictor
Add support for user input, allowing fans to search and track any artist
Integrate more Spotify metrics (e.g., playlists, albums, recent growth trends)
Include a recommendation system based on genre and popularity
Enable users to save favorite artists and see predictions over time
Explore machine learning models to improve prediction accuracy