Recently, I attended a hackathon of the AI Society at KTH Royal Institute of Technology in Stockholm, Sweden. The boundaries of the topic choice were quite loosely defined and the organisers encouraged us to use artificial intelligence (AI). My eventual development can be found on DevPost.
My team consisted out of two other developers and me. All of us had never developed any AI technology before. After an intensive brainstorming, we decided on a use case of analysing travel data with some machine learning algorithms. After several setbacks, including a crashed operating system which took up a significant amount of the hackathon time to work around, I decided to tackle our use case from another perspective. I programmed the whole night and came up with Traap – the TRAvel mAP.
Hackathon Idea: Traap
The main idea of Traap is to convince the user to provide his/her travel data (at least the travel destinations, arrival dates and departure dates). In return, the user gets a visually appealing map which displays his/her travel data in a new way. The user will be able to share these data in existing social networks and allow others to enjoy his/her past trips. A more detailed explanation of this hackathon project can be found on DevPost.
I programmed Traap in Java with the frameworks Spring Boot and Vaadin Flow. Using these frameworks, I built a web backend and frontend in Java, learnt about the possibility to utilise the library Leaflet for maps and tried the OpenCage Geocoder API as an alternative to the Google Maps Geocoding API.
Even though I didn’t exactly pick up some new AI technologies, I did learn a lot about the speedy setup of a web application. The organisers of the hackathon, who are members of the KTH AI Society, liked this idea of Traap and the implementation of the data collection very much and awarded this project a special prize, which rewarded my sleepless learning experience.
If you want to read about another hackathon I attended, have look at the article about the Fiction2Science Hackathon.