Flock Finder

Inspiration

Our interpretation of the theme, "Happiness Through Hobbies", was that we needed to build an app to allow people of a specific hobby to be able to expand their scope. We settled on bird watching because it is done outdoors, is very relaxing and low maintenance, and is popular world-wide. A study done by the NIH proved that therapeutic, outdoor activities done in a group setting had the highest positive effect on mental health in the participants. The study tested many types of outdoor activities, tracked six different mental health markers, and categorized participants into individual and group settings to come to this conclusion. Bird watching fits into the category of "therapeutic, outdoor activity", the only thing that could maximize the happiness of the hobbyists is something that can connect the global population together!

What it does

Our app allows users to photograph and identify birds that they find, upload the post to a global 3D map, and navigate the map in search of posts by other users. Users can explore the map to find rare birds, local birds, or even birds from around the world that they may want to travel and see. If our app identifies a user's bird sighting as rare for the area, the user will gain points and achievements.

How we built it

Using a hugging face dataset, we trained ResNetl8 to identify different bird species from images. We have a Google Cloud Run server which runs on SpringBoot. Our Images are stored in a blob database using Google Cloud Storage. Our main SQL database is stored on Google Cloud SQL database. We store the model in Google Cloud Storage, run the model in Google Cloud Run. When the user creates a post, the database writes the image to blob storage, stores metadata of the image to the SQLDatabase, at the same time the image is uploaded to the prediction server to predict the species of bird. The prediction data is then appended to the SQL server. The User Interface comes from the react-native framework.

Spring Boot Backend

Challenges we ran into

We sourced some data from Wikipedia, and trying to figure out the API to do so was difficult. We were unclear about the rules of the contest as some of them changed during the contest. Coming up with an idea for the app was difficult as well.

Accomplishments that we're proud of

We are proud of the server architecture for the app. Additionally, getting the map feature for the frontend to work was rewarding. We are also proud of the logo.