PowerGrader

Grader UI Demo!

SODA Hackathon Proof of Concept

Project: PowerGrader

Team: SSGA

Date: October 26th, 2024

Team Members: Gonzalo Allendes, Saatvik Choudhary, Anadi Dewan, Swayam Mehda

Chosen Track: SODA: Develop software that can scale and improve accessibility to education.

Overview

Project Description:

PowerGrader is a web application for automated grading in written assignments submitted 
to online educational platforms. With the rise of online learning and the widespread use 
of online course management systems, there is an increasing need for providing the
functionality to assess students' learning automatically. Nowadays, available tools 
provide support for multiple choice exams or embedded coding, but they do not support
automated grading for written submissions. While universities may benefit from freeing
educators to focus on high-impact teaching, free online courses may benefit from an 
automated grader like PowerGrader to offer the course administrator the option to 
request written assignments from students. 

The implementation of automated grading for written submissions represents a significant 
advancement in online education, paving the way for its large scale application in 
university courses. This shift could lead to improved cost efficiency in grading tasks,
allowing educators to allocate more time to instructional activities and student 
engagement.

Project Presentation

Link to Student UI Video Demonstration

Link to Grader UI Video Demonstration

Link to Github Repo

Tech Stack

Next.js: This project was created with the Next.JS React framework for building full-stack applications.

PostgreSQL: This project uses PostgreSQL to integrate a database that manages user roles and submission, information, along with grades and paths to the stored files.

Firebase: Firebase serves as the file storage system for assignment submissions.

Next-auth: We utilize NextAuth to enable secure session management, supporting Google authentication.

Gemini: This project integrates Google’s Generative AI framework, Gemini, to analyze files and generate responses based on a grading rubric.

Flask: Flask is integrated into the project to facilitate API calls to the database.

Project Plan

Key Requirements:

Requirement	Description
User Authentication	Course administrators, students, and graders must be able to log in to the platform.
Assignment Submission	Students should be able to submit written assignments in various formats (e.g., PDF, DOCX).
Automated Grading Engine	The application must implement an AI-powered algorithm to grade assignments.
Normal Distribution	The algorithm should apply a grading curve after processing all submissions.
Feedback Mechanism	Users should receive automated feedback based on grading criteria, including suggestions for improvement.
Customizable Rubrics	Instructors should be able to define specific grading rubrics and criteria for different assignments.
Analytics Dashboard	Provide an interface for instructors to view analytics on student performance and assignment submissions.