StudyBuddy

interface

Inspiration

The main inspiration comes from the emerging age of a new way of thinking of UIs which involve users querying an AI for the information that they're looking for rather than scrolling through a traditional UI to find information. We also wanted to build something with a strong foundation that could grow into something so much more.

What it does

Study Buddy is a student's personal guide to their class's documents and past assignments. When you select a past assignment, you can get personalized tutoring based on which questions you got wrong and which ones were answered correctly. We have pre-prompted the assignment tutoring to allow student's to have a variety of options from receiving help on specific questions, or on grasping the main points of the assignment all-together.

Study Buddy accesses your course documents in order to get context on what you're being taught so that it can know the best way to personally tutor you. This feature is also extremely beneficial for refreshing on material before an exam or learning more about a topic presented during a lecture.

How we built it

The front end was created using TypeScript/React. We are currently using a flask app as a temporary backend to make API calls for the Canvas API, LLM API, and Documents Loaders/Transformers. We used the Langchain library for the LLM Prompt engineering and chaining.

Challenges we ran into

A big challenge early on was navigating the canvas API. Considering no one in the group had used the Canvas API before, it was difficult to learn how to pull specific elements from our Canvas accounts to the correct location in our app. Our poor initial planning also showed to be costly as we used an incompatible tech stack at times that cumulatively set us back a few hours. Accomplishments that we're proud of We're very proud that we were able to create a product from scratch that us, along with our friends find very valuable. Our group built up a level of resilience that some of us hadn't known before this weekend and I believe many of us are proud of completing this project in the end.

What we learned

We learned that planning every element of a project should not be overlooked. We now know how important it is to completely understand your data flows and tech stack for the project before you begin. This would've saved us a bunch of time in the long run.

What's next for StudyBuddy

We need to fix a few bugs to make the app fully functional. Beyond that, next in store for StudyBuddy is to be able to automatically build a knowledge base with all of the files in your canvas courses (PowerPoints, PDFs, lecture transcripts, syllabus) that you can then ask the chat bot about. We also intend to optimize the UI, as the current view is just a general markup of what we want.