Divya Patel

Projects

A curated collection of my latest projects

10, Dec 2023

Analysis and Visualization of Financial Health of Residents

I spearheaded the development of advanced D3.js visualizations as part of a group project focused on analyzing economic trends in Engagement, Ohio, over a 15-month period, based on data from the VAST MC 2022 Challenge 3. Our interactive system leveraged diverse visualization techniques, including an interactive city map, line and bar charts, Sankey diagrams, and scatter plots, to provide a comprehensive understanding of business dynamics, resident financial health, and employee retention. My contributions significantly enhanced user interactions, enabling more intuitive exploration and insightful analysis of complex datasets.

15, May 2024

Scalable Video Analysis with AWS Lambda

In this project, I developed a robust and scalable cloud-based video analysis application utilizing AWS Lambda and supporting AWS services. The application was designed to automatically scale in response to demand, providing a cost-effective solution for processing video uploads. The system implemented a multi-stage pipeline, where each stage was managed by separate Lambda functions responsible for tasks such as video splitting, motion detection, face extraction, and face recognition. This project demonstrated my ability to create an advanced serverless application, leveraging the power of PaaS to build a highly efficient and scalable cloud service.

15, Nov 2022

An Effective EEG Signal-Based Sleep Staging System using Machine Learning Techniques

In this project, I developed an automated sleep staging system that utilized machine learning techniques to classify EEG signals into five distinct sleep stages: Wake, S1, S2, S3, and S4. Leveraging the PhysioNet dataset, the project involved extensive data preprocessing and feature extraction from EEG signals. I implemented and compared multiple classifiers, including KNN, RUSBoost, and Random Forest, to evaluate their performance. The Random Forest classifier achieved an impressive accuracy of 96%, significantly outperforming other models. This work demonstrates my ability to apply advanced machine learning algorithms to real-world biomedical data, contributing to the development of efficient and accurate sleep disorder diagnostics.

15, Mar 2024

Elastic Face Recognition Application on AWS

I developed an elastic face recognition application on Amazon Web Services (AWS), utilizing Infrastructure as a Service (IaaS) resources. In the first phase, I built the web tier on a single EC2 micro instance, which handled image uploads and performed face recognition using a lookup table. This phase laid the groundwork for a scalable and responsive cloud service. In the second phase, I expanded the application by implementing the App Tier with a deep learning model for accurate face recognition, and the Data Tier for persistent storage using S3 buckets. The application featured dynamic autoscaling, efficiently managing up to 20 instances based on demand and processing high volumes of concurrent requests.