Kumar Viswanath Rameswaram

Air pollution is a growing public health concern, and understanding what drives it is key to building smarter, cleaner cities. This project explores the relationship between weather conditions and air quality across five major U.S. cities, namely New York, Los Angeles, Houston, Phoenix, and Denver. Using real-world data pulled from the OpenMeteo API and the U.S. Environmental Protection Agency's Air Quality System, the study focuses on predicting PM2.5 concentrations, a fine particulate matter that is directly harmful to human health. The analysis began with exploratory data analysis to uncover correlations between weather variables like temperature, humidity, wind speed, and pollution levels. While initial machine learning models using Random Forest and Logistic Regression were tested for classifying air quality categories, the project evolved into a time series forecasting approach, using engineered features such as lag values and rolling averages to predict the next day's PM2.5 levels in Denver. The findings demonstrate that weather patterns carry meaningful predictive power for air pollution, offering a practical foundation for early warning systems and data-driven environmental planning.

This project is about building a web tool that helps Amazon sellers quickly understand who their competitors are and how they compare price, ratings, and features. When a seller types in a product ID, the tool automatically finds similar competing products on Amazon, collects their details, and uses an AI model to write a report covering topics like pricing position, product strengths, and growth opportunities. The whole process completes in well under a minute. The tool was built with a Python web server on the backend, a React website on the frontend, and a database to store results. The AI part uses a large language model called Llama running on fast Groq hardware. Testing showed the tool consistently finds a good number of competitor products per search and produces clear, useful reports across fifteen different analysis topics.