Spotify Product Analytics & Recommendation System

This project is a data analytics and machine learning dashboard designed to analyze how different audio features influence song performance and user engagement. It also includes a recommendation system that suggests songs based on similarity and user preferences.

The goal of this project is to apply core data science concepts such as feature engineering, clustering, and recommendation systems on a real-world dataset, while presenting the results through an interactive dashboard.

Live Application

You can explore the deployed application here: https://spotify-interactive-analytic.streamlit.app/

What this project does

Analyzes music data to understand what drives song engagement
Segments songs into meaningful groups using clustering
Recommends similar songs using a content-based approach
Provides simple user-based personalization
Displays insights through an interactive dashboard

Key Features

Analytics Dashboard

Custom engagement score built using popularity, energy, danceability, and duration
Key metrics such as average popularity, engagement, and duration
Interactive filtering for better data exploration

Clustering (Song Segmentation)

Songs are grouped into three clusters using KMeans:

Chill
Party
Focus

This helps identify patterns in different types of music.

Recommendation System

Similar Song Recommendation

Uses cosine similarity on audio features
Suggests songs similar to a selected track

Fuzzy Search

Handles typos in user input

Personalized Recommendations

Builds a user preference profile based on selected songs
Recommends tracks aligned with user taste

A/B Style Insight

Compares high-energy vs low-energy songs
Measures difference in engagement
Uses statistical testing to check significance

Key Insights

High-energy songs tend to achieve higher engagement
Danceable tracks generally perform better
Shorter songs show slightly better performance trends
A large portion of songs fall into the medium engagement range

Tech Stack

Python
Pandas, NumPy
Scikit-learn (clustering, scaling, similarity)
Streamlit (dashboard)
Plotly and Seaborn (visualization)
SciPy (statistical testing)

How to Run

git clone https://github.com/adityamsr2606/spotify-product-analytics.git
cd spotify-product-analytics
pip install -r requirements.txt
streamlit run dashboard/app.py

Motivation

This project was built to move beyond theoretical learning and apply data analytics and machine learning techniques to a practical use case. It focuses on understanding user engagement and building a working recommendation system within an interactive application.

Future Improvements

Enhance recommendation system using hybrid or deep learning approaches
Add playlist generation functionality
Incorporate real-time user interaction data
Integrate with Spotify API

Author

Aditya Mohan Srivastava