🎶 Follower-Based Song Popularity Prediction

This project predicts the popularity of songs on Spotify using machine learning algorithms. By analyzing various song attributes such as artist popularity, follower count, danceability, energy, and more, the model provides valuable insights for artists, producers, and record labels to make informed decisions about song releases and marketing strategies.

---

📜 Table of Contents

• Introduction
• Problem Statement
• Approach
• Significance
• Data
• Technologies Used
• Proposed Models
• Challenges and Learnings
• Future Work
• How to Run the Project
• Contributors
• References

---

🌟 Introduction

The "Follower-Based Song Popularity Prediction" project aims to predict song popularity on Spotify by using machine learning techniques and various song characteristics. Our model uses features like artist popularity, follower count, danceability, energy, acousticness, and more to predict which songs are likely to become hits. We explore multiple machine learning algorithms, including Decision Trees and K-Nearest Neighbors (KNN), to identify the most accurate model for song popularity prediction.

This project provides essential insights into the music industry, offering a way for artists and producers to enhance marketing strategies, optimize release schedules, and maximize listener engagement.

---

🔍 Problem Statement

Predicting song popularity is a challenging task, as many factors contribute to whether a song becomes a hit. In this project, we address the following problem:

• Goal: To predict song popularity using a comprehensive set of features such as artist popularity, follower count, and song attributes (danceability, energy, loudness, acousticness, etc.).
• Objective: Provide actionable insights to artists, producers, and record labels to optimize marketing efforts and release strategies.

---

💡 Approach

The project follows a structured approach to predict song popularity:

1. Data Collection: Spotify API is used to gather data on song attributes such as danceability, energy, loudness, and artist details like popularity and follower count.
2. Data Preprocessing: Raw data is cleaned, normalized, and processed to handle missing values and irrelevant features. This includes selecting relevant features, removing duplicates, and standardizing numerical and categorical data.
3. Model Development: Machine learning algorithms such as Decision Trees and KNN are trained on the preprocessed dataset. The models are optimized through cross-validation, hyperparameter tuning, and performance evaluation.
4. Model Comparison: The performance of each model is evaluated using metrics like Root Mean Squared Error (RMSE), and the most accurate model is selected for song popularity prediction.

---

🌍 Significance

This project offers significant value to the music industry by providing predictive insights that can influence marketing and production strategies. The model helps:

• Artists & Producers: Optimize release schedules and marketing campaigns by predicting which songs are likely to become hits.
• Record Labels: Make data-driven decisions on resource allocation for promoting specific songs or artists.
• Global Scalability: The model can be adapted to predict song popularity across different genres, regions, and markets, making it a valuable tool for global music analytics.

---

📊 Data

a. Data Gathering

• Data is collected from Spotify’s API, focusing on features like artist popularity, follower count, and various song attributes.
• This data is crucial for understanding the dynamics of song popularity.

b. Data Preparation

• Data cleaning involves removing irrelevant features, handling missing values, and ensuring consistent data formats.
• Numerical features such as danceability and energy are normalized, and categorical features such as artist country are standardized.

c. Data Preprocessing

• Features like track ID and song name are removed as they do not contribute to prediction accuracy.
• Duplicates are handled, and a dataset without null values is created for the model development phase.

---

🛠 Technologies Used

• Python: The primary programming language for the project.
• Pandas & NumPy: For data manipulation and numerical operations.
• Matplotlib & Seaborn: For creating visualizations to explore data relationships.
• Scikit-learn: Machine learning library used for training and testing the models.
• Google Colab: The development environment where the models are trained and evaluated.

---

📈 Proposed Models

1. K-Nearest Neighbors (KNN)

• The KNN algorithm is applied to predict song popularity based on its similarity to other songs.
• The model's performance is optimized by adjusting the number of neighbors and evaluated using RMSE (Root Mean Squared Error).

2. Decision Tree Regressor

• A Decision Tree model is trained to predict song popularity based on features like danceability, energy, and acousticness.
• The model's complexity is controlled through hyperparameters like max_leaf_nodes and min_samples_split, and its performance is measured using RMSE.

---

🧠 Challenges and Learnings

• Data Cleaning: Handling inconsistencies, missing values, and irrelevant features was critical for ensuring data quality.
• Feature Selection: Determining which features contributed most to predicting song popularity was a key aspect of model development.
• Model Optimization: Tuning the hyperparameters of the models to balance accuracy and overfitting was challenging but essential for improving performance.

---

🔮 Future Work

• Advanced Algorithms: Incorporating deep learning models to improve the accuracy of predictions.
• Real-Time Predictions: Expanding the model to predict song popularity in real-time.
• Broader Data Integration: Incorporating additional data sources, such as social media sentiment and concert ticket sales, to enhance the model’s predictive power.

---

🛠 How to Run the Project

1. Clone the Repository:

   git clone https://github.com/your-username/follower-based-song-popularity-prediction.git
   cd follower-based-song-popularity-prediction
   

2. Install the required libraries listed in the requirements.txt file. pip install -r requirements.txt

3. Run the Jupyter Notebook: Open the popularity_prediction.ipynb notebook in Google Colab or locally, and run all the cells to train and test the models.

4. View Results: The predictions and visualizations will be displayed after running the notebook, including charts and metrics for model performance.

---

📚 References

1. "SpotHitPy: A Study For ML-Based Song Hit Prediction Using Spotify"
2. Revisiting the Problem of Audio-based Hit Song Prediction using Convolutional Neural Networks
3. A. H. Raza and K. Nanath, “Predicting a hit song with machine learning: Is there an apriori secret formula?”
4. E. Zangerle, M. Vötter, R. Huber, and Y.-H. Yang, “Hit song prediction: Leveraging low- and high-level audio features”