Projects
Customer Retention and Business Intelligence Analysis
Developed comprehensive Power BI reports and Python analytics to analyze customer retention patterns across product categories, identifying key drivers of 15% higher retention in accessories and optimizing profit margins (40-50%) in bike segments.
Earthquake Forecasting
Implemented advanced machine learning algorithms for seismic events, enhancing model precision by 30% and contributing to early warning systems and disaster management strategies.​
Walmart Sales Prediction
Developed and deployed machine learning models to forecast Walmart sales trends, incorporating feature selection techniques and performance metrics to enhance real-time recommender systems and decision support.
Opioid Prescription Analysis using Machine Learning
Generated predictive models using decision trees with hyperparameter tuning and logistic regression to identify over 1,000 potential high-risk prescribers by interpreting prescription patterns.
Prediction of Agricultural N2O Emissions
Applied diverse machine learning techniques to predict N2O flux, utilizing forward selection and backward elimination for optimal feature set identification. This approach enhanced model accuracy and interpretability for environmental data analysis.
Breast Cancer Detection using Deep Learning
Developed an Android application using a Convolutional Neural Network (CNN) to classify breast cancer as malignant or benign from medical images. This tool aims to enhance early detection and improve accessibility of cancer screening.
IOT based Accident Prone Device
Engineered a compact motorcycle-mounted device that instantly sends accident notifications, along with the user's location, to emergency services. The same technology was extended to CCTV cameras through TensorFlow, enabling real-time accident detection and alerting authorities.
