Visit our app: https://ctpfallgroupproject-vibify.streamlit.app/
Vibify, short for "Vibe-Identify," is a web application designed to support the songwriting process. Leveraging NLP (Natural Language Processing) and deep learning, this tool analyzes lyrical data, identifies its sentiment, and offers inspiration through lyrically similar songs.
Vibify is tailored for songwriters, composers, and anyone intrigued by the art of songwriting. It assists in:
- Assessing the emotional valence of lyrics, providing a spectrum from sadness to happiness
- Overcoming writer's block by suggesting similar songs for inspiration
- Sentiment Analysis: Using TensorFlow and NLTK, Vibify accurately classifies lyrics into positive or negative sentiment categories
- Genre Detection: It determines the genre of the lyrics, including Pop, Rock, R&B, Country, etc
- Song Recommendations: Vibify suggests the top-5 lyrically similar songs, aiding in the creative process
- Frameworks and Libraries: Streamlit, TensorFlow, NLTK, Scikit-learn
- Data Sources: A comprehensive dataset from Kaggle and additional features from Genius.com
- Hosting: Hosted on Streamlit.io, ensuring seamless user experience and accessibility
- User Interface: The app features a user-friendly navbar for easy navigation
- Dashboard: Upon input submission, the dashboard dynamically displays the analysis results
- Reset Functionality: Users can reset the dashboard to analyze new inputs
- Developed during CUNY Tech Prep's Fall 2022 cohort
- Managed using Trello for task tracking and GitHub for version control
- Regular team meetings and collaboration via Zoom and Slack
- Aleksandra Georgievska @aleksgeorgi
- Deepankar Chakraborty @deepankarnk2
- Stephen Williams @svalentinow
- Dataset Expansion: To include a wider range of songs and genres
- Enhanced User Input Options: Allowing genre-specific song recommendations
- Audio Analysis: Incorporating audio data for a more holistic sentiment analysis
Discover how Vibify can elevate your songwriting experience: Vibify App
- https://www.kaggle.com/datasets/edenbd/150k-lyrics-labeled-with-spotify-valence
- https://www.kaggle.com/datasets/saurabhshahane/spotgen-music-dataset
- Additional featurs were added by webscraping Genius.com