Description

Analyzed Spotify user behavior from a structured survey of 522 individuals to identify churn risk and predict premium subscription intent using machine learning and explainable AI techniques.

The problem

Understanding user behavior to reduce churn and increase premium subscriptions through data-driven strategies for Spotify to market themselves in India and create a monopoly over JioSavan, Apple Music and other competitors.

Key skills used

• Classification modeling (Logistic Regression, Random Forest, XGBoost)
• Exploratory Data Analysis (EDA) and feature engineering
• SHAP-based model interpretability

Some Visualizations

Figure 1: Bar plots from exploratory data analysis showing interest in Spotify Premium categorized by age group, device type, and membership retention willingness

Figure 2 : A Correlation heatmap obtained during EDA which identifies strong relations between numeric columns

Figure 3 : Heatmap obtained after churn analysis which categorizes risk segments based on the plan type