Hey, I'm Shashank

Emonet - Sentiment Analysis but Good

Emonet leverages Gated Recurrent Neural Networks (GRNN) to delve into the complex realm of emotions and feelings. Artificial Intelligence (AI) often struggles with the intricacies of understanding emotions, posing a significant challenge. Emonet aims to overcome this barrier by effectively training AI to comprehend and recognize emotions, introducing a new dimension to the capabilities of machine learning.

Emonet builds a diverse dataset using hashtags, providing a unique alternative to manual annotation. The GRNN model, based on psychological studies, enhances the understanding of emotions in AI. The project primarily works with the LSTM model and compares it with online classifiers to test performance. Emonet achieves an impressive 83% average score in determining emotions, outperforming sentiment analysis models that typically achieve 65-70%. The project's modeling architecture uses a dataset of 20,000 tweets, employing GRNN variations for enhanced performance.

Close Window

Gitlab Analyzer

GitLab Analyzer is a comprehensive analysis tool designed to scrutinize user contributions within a GitLab project. Tailored for the needs of instructors and teaching assistants, this software proves invaluable for efficiently grading students' contributions to team projects. Its functionality extends across diverse metrics, encompassing aspects such as code contributions through commits and merge requests, GitLab issue interactions, and feedback from code review comments. By offering a detailed and multifaceted examination of user engagement, GitLab Analyzer empowers educators to assess and evaluate the depth and quality of students' involvement in their collaborative endeavors.

This piece of software provides a centralized platform for instructors to gain insights into various facets of student participation, facilitating a more nuanced and informed grading process. Whether evaluating the frequency and significance of code contributions or gauging the level of engagement through issue resolution and code review feedback, GitLab Analyzer streamlines the grading workflow. By leveraging these metrics, educators can ensure a fair and comprehensive assessment of each student's involvement in the project, fostering a more objective and insightful approach to grading collaborative assignments.

Close Window

Community Based Rehabilitation

The Community Based Rehabilitation project is an Android application developed for the London-based NGO, Hope Health Action, with a specific focus on assisting Ugandan refugees facing disabilities and health challenges. The primary objective of this initiative was to digitize and overhaul the entire database system, aiming to transform the process of data collection and storage for increased efficiency and ease of use by field workers.

This comprehensive project involved the creation of a backend web server complemented by a frontend Android application. Notably, the Android app serves as a direct communication channel connecting office workers in London with field personnel, facilitating announcements, critical messages, and daily bulletin updates.

Close Window

MySchedule Difficulty Calculator

Why spend valuable time delving into extensive research on courses and professors when there's an application designed to handle it all for you? The Workload Calculator is a comprehensive tool that streamlines the pre-enrollment research process for students by consolidating all necessary information into a single, user-friendly environment. Leveraging the Coursys API, it selectively displays relevant courses based on the chosen semester, while also extracting and refining data from RateMyProf and CourseDiggers. The application goes a step further by presenting an approximate weekly workload and an overall difficulty score for each course.

Close Window

Submarine

Lorem ipsum dolor sit amet, consectetur adipisicing elit. Mollitia neque assumenda ipsam nihil, molestias magnam, recusandae quos quis inventore quisquam velit asperiores, vitae? Reprehenderit soluta, eos quod consequuntur itaque. Nam.

Close Window