Hatch | Devpost

Hatch

Inspiration

This api was a project for HackUPC 2022. The aim of the project was to create a game similar to GeoCatching but inside the venue of the Hack and coperative between all people to make hackathons even funnier! This api is easily integrable with any frontend. It has been developed using GraphQL so it makes even easyer the integration with front techs as Data can be queryed on-demand. Eventhough we had not been on time to present you the API with a client, we want to let you know what is behind Hatch.

What it does

As it has been said, this API is all the back logic of a Game called Hatch. Staff and administrators of the hackathon will be able to prepare a "Session" in which they will be able to register some "Caches" that hackers will have to find. Those Chaches are especific objects, places that can be dispersed through all the venue.

This is a colaborative game, this means that ALL HACKERS will be collaborating to find all the Caches that the Session contains. If one single Hacker finds one cache, all other hackers will go through the next level.

Once a Cache is found, you can report this through a photo to the backend. There, we have an algorithm that computes the Image Similarity between the one updated by the administrator once created the Cache and the new submmited by the participant.

If the image is similar, which means that is the Cache, that cache will be marked as reached and the next one will appear.

Each cache has a determined the number of clues that are incrementally going public as the time without finding one passes.

How we built it

Api entarelly build with Graphene-Django, a GraphQL library.

Challenges we ran into

Learning new Technologies such as GeraphQL and React Native, even we didn't came out with a client solution. Also working with AI and images to work with image similarity.

Accomplishments that we're proud of

Good API easy to integrate for all organizations that are organizing Hackathons and would like to make there Hackathon even funnier. Coming out with a good enough image similarity AI Algorithm to discriminate objects.