EduAdvisor AI

Inspiration

Being Students at IU, we have experienced the stresses of scheduling classes and having unresponsive advisors. We felt inspired by helpful chatbots and AI tools like ChatGPT and wanted to deliver a personalized and helpful experience to the user just like other agents.

What it does

Our program works by hosting a dialogue with the user. The user is able to type in their interests, their major, what semester they are in, and other personalized information. The AI chat bot is then able to recognize what it needs to offer suggested majors based on interests if they would like and provide a specialized course schedule for the student. The schedule ensures that no times are overlapping, that the required courses are taken, and that gen-eds recommended by interest are included. The AI chatbot further provides a visual diagram showing how the user's interests are related to their courses.

How we built it

We started off by going to the drawing board and planning out what we wanted to do. From there we went and scraped data from multiple IU websites, gathering 100's of thousands of rows of relevant data. From there we went and cleaned the data using Python and Pandas. We quickly made a frontend program with Vite, React, and Tailwind to act as our UI. We linked the two parts by hosting our backend on a PythonAnywhere server Gabe had, using http requests to exchange data and secure our API. Utilizing our data with the Gemini API, we were able to leverage natural language processing with real results to help users gain real answers related to IU. We then deployed our frontend with github pages, making it ready for you to use!

Challenges we ran into

For many of us, this was the first time we worked in a team for a coding application. Many of us have little to no experience with using GitHub in a team environment. This caused issues with pulling and pushing and sharing the right information. As the process went on, we did start to see more success. Another challenge we ran into was the formatting of the data, as none of the data was in a very clean format. We had to use python and pandas and took quite a long time to get readable csv's of the data.

Accomplishments that we're proud of

We are beyond proud of our working demo that efficiently shows how our project can problem solve and offer class scheduling and recommendation solutions to students.