Undertow

Undertow connects us with the past while creating genuine connections in the present.

Comment 1

Inspiration

We are inspired by the increasing lack of genuine connection in the modern day. As social media and online entertainment becomes more and more prevalent, people become more inclined to stay at home, and less inclined to go out and meet real people. On a more personal level, the USYD campus is one that is abundant in heritage. As thousands of students walk in and out of these buildings every day, none of them are aware of the rich history and culture that these buildings hold. This is why we created Undertow, demonstrating the connecting between the past and the present, and garnering genuine connections with people nearby.

What it does

After signing up with your real name and phone number, with an optional profile picture, Undertow scans an image and your location to determine a landmark you are exploring. After scanning, you are put into the present, where you can see comments from the people who have visited the same landmark within the past hour. Each comment has a number of likes, how long ago it was posted, and always has a number to contact, which will copy the persons number to the clipboard. This fosters a phone call conversation from a comment with someone who would not be geographically far away as the comment is recent.

By pressing the info button, you also get a short overview of the building's significance and its history. You are also able to post your own comment, with fields for what you are doing at the time, a short description, and even a photo if you wish. The possibilities to express your personality through the app are very broad.

Swiping right takes you to the past, with an AR simulation of what the landmark used to look like, and holograms of past historical figures, doing what they used to do. Holograms have descriptions that can be clicked to learn more about the history and how things worked in the past. There are also chosen highlight comments in the past, typically from longer than 5 years ago, that further illustrate what was happening in the past.

How we built it

We spent a long time figuring out Github, then assigning tasks across the team after the initial ideation. We decided on html css for the front-end, javascript for the backend, and did majority of the coding through Claude Code and Visual Studio Code, linked to the repository. We had one man each working on separate branches for the front-end, back-end, database, and AI-detection. Two of the members synthesised everything to a more polished product.

For the location detection, we identified the longitude and latitude of each significant building and compared it to the users existing location, implementing Harversine's formula to ensure accuracy and reliability. The final app would ideally have an AR based landmark recognition system.

Each landmark location has a well-researched summary, that can be accessed through the info button, that is stored in a database. This is the same with descriptions for the historical figure holograms. Ideally in the final product we would have a full AR simulation with holograms, but our MVP uses an image with gyroscopic movement matching the phone's movement with a hologram animation.

Our final MVP uses Netlify to create a usable web app.

Challenges we ran into

As this was the first hackathon for all of us, Github was a massive challenge to get familiarised with. The first four hours of hacking was almost exclusively spent on figuring out how to connect to a repository and update our branches and main.

During the coding, we were faced with the challenges of imprecise location tracking adversely affecting the detection of the building in the USYD campus. We mitigated this issue through editing the formulas around the location tracking, but the AR recognition in the proposed final product would prevent this issue.

The front-end that was coded by one member did not initially work with the back-end created by another member. One group member took the time to synthesise the ideas to a functional prototype. Our back-end also initially was not able to store our comments, and likes were not functional. After a bit of tinkering, we managed to make comments functional on local machines, that is, it would remember comments made on the same laptop. However, this was eventually massively improved on, with server-side storing of comments, time-tracking and improved like counters/animations.

The final notable issue was with the holograms. We were very ambitious in hoping we could create holograms in an AR environment, but this proved too challenging for a 24 hour hackathon. We resulting using an image with gyroscopic tracking, and an animation that is stuck on top of the image. This also made it quite simple to add the info box about the person.

Accomplishments that we're proud of

We are especially proud of the working Github repository, as this took up a large amount of our hacking time. It was, however, found later than AI could guide us through most of it once the base understanding is there for it.

We are also very proud of the working location-landmark tracking system within USYD. Our app is able to recognise main landmark buildings on campus such as J12 Computer Science, Seymour Centre and Charles Perkins Centre.

Furthermore, we are proud that we made a working and engaging comment system that allows people to show their personality, connect and reflect with each other.

Finally, we are exceedingly proud that we were able to create a viable product that properly envisions our initial idea in our very first hackathon.

What we learned

We learned:

  • Github Repo's and Branches
  • Javascript
  • VS Code
  • Netlify
  • Location Identification
  • Possibilities of Modern Technology
  • The History of USYD Buildings

What's next for Undertow

Undertow seeks to expand it's landmark database far beyond just the current USYD campus. This would then allow people all over Sydney to connect with each other who visit the same landmarks and have the same shared experiences.

The step after that would be to integrate Undertow into mobile app stores so that people can actually download the app. Of course there would need to be significant improvements to the back-end to ensure this process goes smoothly and higher numbers of people use the app.

The final, much bigger hurdle is to fully develop the AR-nature of the app. We would eventually like for the app to not take a photo to recognise the building, but for it to recognise the building as you pan past, which clicking will take you into the present. Moreover, we would like there to be holograms in the present that represent how many people have visited recently. This AR-nature should also be developed in the past, where what the building used to look like should be overlayed to look like it replaced the current building. It is a similar idea with the holograms of historical figures.

Built With

Share this project:

Updates