HackStack

Project Info

Winner

Education Hacks; [MLH] Best Use of Auth0

Tide Party

Devpost

Inspiration

Hey! We're TideParty, but our names are Oleg and Anthony. We LOVE tide pooling in Santa Cruz. It’s a relaxing activity that many people enjoy, and Santa Cruz is an ideal place for it thanks to its rocky cliffs and abundant marine life. However, when people go tide pooling, they often don’t know what they’re looking at. We built this app to help people identify marine organisms in real time, learn about them, and gain a deeper appreciation for ocean conservation.

What it does

Tide Party is an interactive mobile app that turns tide pooling into a fun, educational experience: Discover and navigate to local tide pooling spots Identify marine life in real time by pointing your camera at a sea creature Take photos and collect the creatures you find Learn detailed information about each species Earn badges by collecting creatures and answering quiz questions Join party rooms with friends and race to find the most creatures

How we built it

We built the app using SwiftUI for the frontend and Firebase for backend storage and real-time features. For marine life recognition, we trained a custom machine learning model using PyTorch and converted it into an Apple-compatible Core ML model for on-device inference.

Challenges we ran into

Training the model was our biggest challenge. Due to the limited available image data, we relied heavily on image augmentation to improve performance. Collecting and labeling data was also difficult, and early versions of the model suffered from underfitting. After extensive preprocessing and training, we were able to build a generalized model capable of accurately classifying Santa Cruz tide pool marine life with 94% accuracy.

Accomplishments we're proud of

We’re proud that we were able to fully realize our original vision, from real-time classification to multiplayer features. Achieving high model accuracy, building a polished UI, and designing a friendly mascot to enhance the user experience were major wins for us.

What we learned

We learned how to work quickly and effectively under pressure, how to design thoughtful UI/UX animations and interactions in Swift, and how to train and deploy a machine learning model within a real mobile application.

What's next

Next, we want to release Tide Party to real users. It would be especially impactful for school field trips, where students can compete to discover marine life together. We also plan to: Expand and improve our tide pool location map Train a more robust model to classify additional species Support multiple creature detections in a single frame Add online leaderboards and public party rooms Partner with charities focused on local ocean conservation

Analysis

Compare with all teams

View

Metric

Figures cover GitHub contributors during the hackathon window. A co-authored commit counts in full for each author, so per-member totals add up to more than the whole-team figures.

Technology

Found in codeClaimed only
  • SwiftIn code
  • FirebaseClaimed
  • Google GeminiClaimed
  • PythonClaimed
  • PyTorchClaimed

1 of 5 appear in the indexed code. 4 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

306 KB

Source files

52

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

0 stars