Music Recommender System

Overview

This project implements a Music Recommender System using Python, Streamlit, NLTK, and scikit-learn. The system analyzes a Spotify dataset, applies text processing techniques, and calculates cosine similarity to provide song recommendations based on user selections. Users can rate songs, provide feedback, and view personalized recommendations.

Prerequisites

Prior to running the project, you need to download the Spotify dataset (spotify_millsongdata.csv) from Kaggle and place it in the project directory. You can find the dataset here

Features

• Data Preprocessing:

  • Cleans and preprocesses Spotify dataset, including limiting the dataset to the first 50 rows.
  • Converts text data to lowercase, removes specific characters, and performs stemming using NLTK's PorterStemmer.

• TF-IDF Vectorization:

  • Utilizes scikit-learn's TfidfVectorizer to convert processed text data into a TF-IDF matrix.

• Cosine Similarity:

  • Computes cosine similarity between songs based on the TF-IDF matrix.

• Recommendation Function:

  • Provides song recommendations using a precomputed similarity matrix.

• User Feedback and Ratings:

  • Stores user ratings and feedback in dictionaries for data persistence.
  • Implements a user-friendly interface for rating songs and providing optional feedback.

• Streamlit App:

  • Utilizes Streamlit for building a web application with an interactive user interface.
  • Includes features like showing the DataFrame, song selection, rating sliders, and feedback text area.

Usage

1. Install dependencies: pip install -r requirements.txt
2. Run the app: streamlit run app.py

Project Structure

• app.py: Main Streamlit application script.
• user_ratings.pkl: Pickle file for storing user ratings.
• user_feedback.pkl: Pickle file for storing user feedback.
• requirements.txt: List of project dependencies.

Getting Started

1. Clone the repository: git clone https://github.com/grace073/Music-Recommender.git
2. Navigate to the project directory: cd Music-Recommender
3. Import the dataset as spotify_millsongdata.csv from kaggle
4. Install dependencies: pip install -r requirements.txt
5. Run the app: streamlit run app.py