https://anis105.github.io/gtg/

Motivation

Simplifying College Course Selection.

Navigating college academics can be challenging, especially for underrepresented groups. To alleviate this, we created a tool that uses personalized metrics and past data to help students schedule courses. Our tool analyzes feedback and grades, making it easier to plan a successful academic journey.

Dataset

We utilize the PlanetTerp website and its API to access course, professor, and grade data from the University of Maryland. This data is scraped and processed for analysis.

Methodology

We employ Large Language Models (LLMs) for data processing, including Sentiment Analysis and Wordclouds. Sentiment Analysis calculates polarity and subjectivity scores for class reviews. We also developed an algorithm to assess the "danger" of a student's class registration choices.

Application Overview

The deployed version of this application can be accessed at [link_to_your_deployed_application]. It consists of three main parts:

• Class Registration Navigator
• Class Word Cloud
• Class Emotional Heatmap

Solutions / Project Approach

In this React application, we've designed three key components to help students navigate their college course selection:

• Class Registration Navigator:
  • The heart of the application, this component assists students in planning their course schedule effectively. It infuses insights into course feedback and past class review drive by LLM(Large Language Model) and NLP(Natural Language Process), aiding in well-informed decisions.
• Class Word Cloud:
  • This component offers a visual representation of the most commonly used words in course reviews. It helps students quickly grasp the sentiment and focus areas of different classes.
• Class Emotional Heatmap:
  • The heatmap provides a unique perspective on course satisfaction. By visualizing emotional responses from class reviews, students can gauge the emotional tone and overall experience of a particular course. To sum up, By utilizing OpenAI's "gpt-3.5-turbo" model, we performed semantic analysis by generating meaningful responses to course reviews, considering both the content and sentiment expressed in the reviews. This approach enables automating the analysis of a large volume of reviews, making it a powerful tool for extracting insights from textual data.

How and When to Use the Application/Recommendation

• Class Registration Navigator:

  • Utilize this component when planning your college course schedule. Input course names to receive personalized insights and recommendations.

• Class Word Cloud:

  • Use this feature to quickly grasp the sentiment and focus areas of different classes. It's a visual aid for understanding course reviews.

• Class Emotional Heat map:

  • Explore this tool to gauge the emotional tone of course reviews. It provides a unique perspective on course satisfaction.

Deployment

Front end (React)

To run the application locally, git clone the repository to local and stay in the main branch. Then run the following commands to run the app in development mode:

npm install

npm start to run locally or

https://anis105.github.io/gtg/

Backend

We use Python scripts to generate the necessary JSON files utilized in the frontend of our application. No need to worry about running these scripts, as the JSON files have already been pushed to the main branch, enabling the smooth functioning of the web app.

Future dicrections

1. Robust Login Module: Our goal is to design a secure and user-friendly login module. It will not only safeguard user data but also simplify the login process, ensuring a more efficient experience for our users. 2. Instructor Recommendation System: We aspire to introduce a state-of-the-art instructor recommendation system. By employing advanced algorithms, this system will suggest educators based on a user’s preferences, previous interactions, and course history. This will empower students to discover instructors who align with their learning style and objectives. 3. Insightful Report Generation: Our commitment is to furnish our users with comprehensive, insightful reports. These reports will provide a deeper comprehension of their learning progress and performance. They won’t just track academic achievements but also offer valuable insights for continual improvement driven by LLM.

Built With

• chatgpt
• python
• react