Project Info

StudyFinder

This project did not submit a demo video on Devpost.

Inspiration

With our busy schedules at UC Berkeley, we sometimes crave buddies to study with, to grind through homework and practice exams. However, oftentimes we struggle to set up a time and place with people we would genuinely enjoy studying with, which inspires us to create this app that can help students find their perfect match to work through tough content together.

What it does

By entering your own preferences, StudyFinder helps to match students with other students to form the perfect group, taking into account many different preferences.

How we built it

We created the Frontend with Flutter to show questions to the users that is then stored in a Chroma database. This database processes the users' inputs and calculates similarity scores to best match users with other users.

Challenges we ran into

We were struggling with the time crunch and given that we only ended up with 2 people in the group, we had to work really hard overnight to crunch out the program. Our code definitely had some issues too where there were often functions that were jumbled up because of how many files there were.

Accomplishments we're proud of

This is our first time at a Hackathon! We are proud that we were able to get on task quick and work through the various challenges with persistence.

What we learned

We need to manage our time better in splitting up the tasks and figuring out an organization that can help our code look neater and more efficient.

What's next

StudyFinder will implement location that can help students form the most efficient study groups where no one has to walk extra far to meet up with others. Users need to allow location access and we will use Google Maps to create the most optimal path. We would also like to expand it to campuses outside of UC Berkeley, to reach a wider range of people.

Analysis

Compare with all teams

View

Metric

No commits on this project resolved to a GitHub account.

Technology

Found in codeClaimed only

CIn code
C++In code
DartIn code
HTMLIn code
KotlinIn code
PythonIn code
SwiftIn code
DjangoClaimed
Google GeminiClaimed

7 of 9 appear in the indexed code. 2 claimed on Devpost could not be matched to code, which may simply mean the tool leaves no trace in the repository.

AI coding agents

No AI coding agent signals were found in this repository.

Detected from committed agent config files and commit authorship. Absence of a signal is not proof an agent was unused.

Codebase size

Source size

62 KB

Source files

Counts recognized source files only; vendored directories, binaries and lockfiles are excluded, so this is smaller than the repository on disk.

Feature verification

3 Verified4 Code-supported2 Claimed only

Django backend API for preferences and matchingVerified
Built with Django and Python for the backend logic
Claimed on Devposthigh confidence
- Backend/myapp/urls.py:4 — url_patterns route add_preferences/ and explore_similar/ to Django views
- Backend/myapp/views.py:1 — Django JsonResponse-based views implement the API endpoints
Django REST Framework CRUD app (Person/apis)Verified
Built with django, python (general backend/API infrastructure)
Claimed on Devpostmedium confidence
- Backend/apis/models.py:5 — Person model with a DRF ModelViewSet and HyperlinkedModelSerializer, appears to be a separate/earlier scaffold unrelated to the actual study-matching feature (uses 'favoriteBoba' fields), not evidence of the matching feature itself but confirms Django+DRF usage
Frontend built with Flutter collecting user preference questionsVerified
Frontend created with Flutter to show questions to the users
Claimed on Devposthigh confidence
- Frontend/flutter_p1/lib/main.dart:86 — CupertinoTextField and CupertinoSwitch widgets collect a self-description blurb and preference toggles on the Finder and Hoster pages
- Frontend/flutter_p1/pubspec.yaml:1 — Flutter project structure and pubspec confirm this is a Flutter app
Gemini used to compute similarity between usersCode-supported
Built with 'gemini' per the Devpost Built With tags
Claimed on Devpostlow confidence
- Backend/geminiProcessingExample.py:1 — standalone script calls google.generativeai to score similarity between two sample texts, but it is a hardcoded experiment/demo (embedded API key, fixed sample strings, plots a histogram) not wired into any Django view or the matching endpoints
Matching displayed to the user (waiting screen and match reveal screen)Code-supported
StudyFinder matches users and shows them their study group/match
Claimed on Devpostlow confidence
- Frontend/flutter_p1/lib/main.dart:246 — ThirdPage shows a static 'waiting for a match' message
- Frontend/flutter_p1/lib/main.dart:261 — FourthPage displays a hardcoded placeholder 'Your match is <xyz>' rather than a real match fetched from the backend
Preference-based matching using embeddings and similarity scoresCode-supported
StudyFinder helps match students with other students by entering preferences, taking into account many different preferences
Claimed on Devpostmedium confidence
- Backend/myapp/views.py:40 — explore_similar_users_view computes cosine similarity across all other users' embeddings to find the most similar match
- Backend/myapp/views.py:43 — calls chroma_manager.get_user.embedding(user_id), a typo (attribute access instead of method call) that would raise AttributeError, so the endpoint as written cannot run end-to-end
- Backend/myapp/chroma_manager.py:34 — get_user_preferences is redefined to call itself recursively (infinite recursion / shadows the working ids-based version), another bug that would break retrieval
Preferences stored in a Chroma vector databaseCode-supported
The Flutter frontend shows questions to users that are then stored in a Chroma database, which processes inputs and calculates similarity scores
Claimed on Devpostmedium confidence
- Backend/myapp/chroma_manager.py:6 — ChromaManage wraps chromadb.Client() and creates a 'user_preferences' collection with add/query methods
- Backend/myapp/views.py:31 — add_user_preferences_view calls chroma_manager.add_user_preferences(user_id, preferences, embedding), but the underlying method signature is (user_id, embedding, metadata), so the raw preference string is stored as the embedding and the real embedding as metadata, a swapped-argument bug
Frontend wired to backend to submit/retrieve preferencesClaimed only
Implied by 'stored in a Chroma database' and 'match students with other students', i.e. the app end-to-end submits preferences and shows a match
Claimed on Devposthigh confidence
Location-based matching using Google MapsClaimed only
What's next: StudyFinder will implement location so students form efficient study groups, using Google Maps for optimal path
Claimed on Devposthigh confidence

An AI agent derived these features from the project’s Devpost page and readme, then searched the code for each one. Verified features are backed by cited code; claimed-only features had no supporting code, which is not by itself proof a feature is missing.