EDA/Model building and Deployment
Health Data Breaches
Health Data Breaches Visualization
Conduct a comprehensive data analysis leveraging the U.S. Department of Health and Human Services (HHS) Office for Civil Rights portal dataset, spanning October 2009 to September 2021. The aim was to unearth intricate patterns and insights pertaining to data security vulnerabilities prevalent in the healthcare sector, employing advanced data science methodologies and the final results were visualized in Tableau Public.
Adv Data Analysis for Optimizing Automobile Insurance Claims and Premiums
The automobile insurance industry faces challenges in accurately predicting claim amounts and setting fair premiums due to the complexity of factors like customer demographics, vehicle types, and policy types by advanced data analysis techniques in R, such as Principal Component Analysis (PCA), regression models, and Correspondence Analysis and identified key factors influencing insurance claims, premiums, and policy choices to provide actionable insights for insurers to optimize premium structures and tailor policies based on risk profiles. Git
Sentiment-Driven Classification of Amazon Footwear Reviews (NLP)
This project aims to address the specific domain of footwear sold on Amazon, leading e-commerce platform, by performing sentiment analysis and opinion mining on customer reviews. The core objective is to develop a sophisticated analysis framework that can accurately classify customer reviews into three distinct sentiment categories: good, bad, and neutral. Git
Diagnosis of Pediatric Appendicitis using AI (Adv ML)
In this Project we have leveraged advanced ML and AI techniques , specifically Multi-Layer Perceptron Random Forest (RF), Support Vector Machine (SVM), and XGBoost algorithms to enhance the diagnostic process for pediatric appendicitis, The results demonstrated that our models, specifically Random Forest and XGBoost, achieved significant improvements in diagnostic accuracy. Git
Object Detection in Monitoring Systems
This project focuses on enhancing object detection for surveillance systems by optimizing the YOLOv11 model. While the pre-trained YOLO model effectively detects common objects like people, bicycles, and cars,it struggled with detecting specialized items such as weapons. To address this, we annotated a dataset with weapon images using Roboflow, Trained and fine-tuned the YOLO model on a custom dataset Our approach
significantly improved detection performance, achieving higher accuracy,precision, and recall for weapon detection. Git
Spotify Song Recommender System
The Goal of this Project is to Develop algorithms to measure text similarity between Spotify songs lyrics, where we will be vectorizing the text data (converting text into numerical vectors) and then calculating the similarity between these vectors to find the relevance for the recommendation systems and Implement content recommendation systems to suggest relevant Songs based on user’s and song type. Git
