Data Quality Audit & Automation using LLM
Developed an automated system for assessing data quality using Large Language Models such as Azure AI and GPT, which addresses issues such as semantic misalignment, inconsistencies, and anomalies in data for Bosch.
The teams worked on data quality improvement algorithms, an integration framework, performance metrics dashboard and a LLM based data quality audit tool.
"Charging Ahead or Fueling Change? Decoding the Green Future of the United States"
Conducting a comparative evaluation of electric vehicles (EVs) and biofuels to determine the most sustainable and economically viable transportation solution for the U.S. clean energy transition.
Utilizing predictive analytics, lifecycle analysis, and global regulatory comparisons to assess environmental impact, economic feasibility, and policy implications.
Yelp Data Transformation for Business Insights
Architected ETL pipeline for Yelp dataset transformation, converting complex nested JSON structures into normalized SQL tables on Azure SQL Server using Pandas JSON normalization and SQLAlchemy, enabling relational integrity through primary/foreign key implementation.
Executed advanced SQL analytics on restaurant industry data, deriving actionable business intelligence through multi-dimensional querying of customer behavior patterns, operational metrics, and market positioning factors to identify strategic growth opportunities.
Chicago: Crime Analysis
The code performs an analysis of Chicago crime data using machine learning. It fetches crime data from the Chicago Data Portal API, preprocesses it, and applies a Random Forest model to predict the primary type of crime.
The model is evaluated using metrics like accuracy, precision, recall, and F1-score. Finally, a classification report is generated to provide detailed performance insights.
Property Value Prediction using Machine Learning
Developed a Random Forest model in R to predict property values, achieving high accuracy (R2 = 0.87, MAE = $65,520) through data cleaning, feature engineering, and robust evaluaPon.
Delivered acPonable insights by analyzing feature importance, refining predicPons for high-value properPes, and showcasing data driven decision-making.
The Financial and ESG dynamics of the EV Industry
Utilized Wolfram for data analysis and visualization to benchmark financial performance and ESG metrics of EV companies.
Conducted comparative ESG assessments using statistical methods in Wolfram, uncovering critical gaps in governance and social practices among industry leaders like Tesla, GM and Ford.