AdmitTree

Inspiration

Many students find the university application process overwhelming, often struggling to find program-specific requirements and personalized advice in one place. We wanted to create a centralized, data-driven platform that simplifies this journey by matching a student's unique profile with their ideal engineering programs.

What it does

AdmitTree acts as a personalized admissions consultant. It takes a student's grades, extracurricular activities, and interests through an intake form and uses a matching algorithm to rank university programs based on their profile. The platform also features an interactive dashboard that tracks application progress and a friendly chatbot that answers questions using specific university data.

How we built it

Frontend: Developed using React, TypeScript, and Vite Backend: Flask API Data & AI: The chatbot is trained using Digital Ocean Database: Mongo Database Matching Engine: A custom UniversityMatcher service processes student data to generate ranked recommendations.

Challenges we ran into

Data Synthesis: One of our primary hurdles was reconciling unstructured data from university program guides with our structured MongoDB schema.
Scoring Logic: Developing a fair and nuanced ranking system in the UniversityMatcher that balances hard requirements (like grade averages) with qualitative factors (like co-op preferences and extracurricular impact).
Deployment Connectivity: Ensuring stable communication between our Digital Ocean-trained chatbot and our local development environments.

What we learned

Full-Stack Integration: We gained experience in connecting a Vite/React frontend with a Python/Flask backend, specifically in managing asynchronous API calls for real-time recommendations.
AI Grounding: We discovered the nuances of training a chatbot to stay within the context of specific datasets, ensuring that "AdmitTree" provides accurate admission advice rather than generic responses.

What's next for AdmitTree

We plan to implement an OCR-based upload system so students can upload their transcripts directly instead of manually entering grades in the Intake form.
Moving beyond "matching," we aim to incorporate historical admissions data to provide students with a percentage-based "likelihood of admission" score.