My Projects
Related Projects
Spotify Listening Behavior Analysis
Analyzed Spotify listening history to uncover behavioral patterns, usage trends, and artist preferences over time, using temporal and contextual features from user playback data.
Identified listening streaks, time-of-day preferences, platform usage trends, and repeat artist behavior across 5,000+ tracks from personal Spotify history.
Spotify Listening Insights Pipeline
Built a end-to-end ETL pipeline that ingests both Spotify API data and GDPR streaming exports, normalizes it into analytics-ready tables, and orchestrates daily loads and historical backfills via Airflow.
Recieved 79,889 historical plays and daily updates, enabling trend analysis on listening behavior and audio features. Maintained 100% task success over 2 days.
This is a two-parter that bridges data engineering and data science. First I built a Spotify data pipeline using Airflow and PostgreSQL, then used that data to uncover mood trends and genre patterns in personal listening history.
Related Projects
Depression Prediction API
Built and deployed a production-grade REST API for depression classification.
Delivered reliable mental-health inferences with 99.9% uptime and avg. response time < 500 ms
Student Depression Classifier
Built a binary classification model to predict student depression from behavioral data using supervised learning.
Achieved high AUC score while handling a dataset of over 27,000 samples
This is a two-parter, with each piece building on the other to form a complete workflow.
For my Data Science class final, I trained and evaluated multiple classification models with TensorFlow and Scikit-learn using a synthetic dataset. Later, I challenged myself to take it further by deploying the model as a RESTful API using FastAPI and Docker. This let me productionize the pipeline and serve real-time predictions — simulating what an end-to-end machine learning workflow looks like in practice.
Pretty Coins
Built a computer vision tool to detect and analyze coins in high-resolution images using OpenCV.
Successfully identified and measured circular objects with high visual clarity and precision
LeBot, Lebron Chatbot
Built a Lebron-themed chatbot using NLP and custom regex patterns.
Achieved 85% chatbot intent accuracy
Weather App
Built a desktop app to display real-time weather using OpenWeatherMap API.
Delivered a responsive UI with real-time API integration
MCU Box Office Revenue Analysis
Explored box office trends for Marvel films using visualizations and analysis.
Visualized performance across 23 MCU films using 3 revenue metrics
Contacts Debt-Tracker
Created a Tkinter-based app to track personal debts using SQLite.
Tracked 20+ contact entries with user-friendly input and validation
Analysis on Student Behavior
Explored factors influencing student performance through dashboards.
Built an interactive dashboard analyzing habits across 3 student demographics