There were 6 tracks presented to us. We picked the ‘Tech for Good’ track supported by ART-AI partner CIVICA. This required us ‘to design a system that could make a positive difference to citizen experience in the aftermath of COVID-19: life-saving, life-enhancing or life-changing’.
Within 24 hours, we had built an App that we called ‘See’ to help people shop safely, easily and effectively by providing real-time customer volume, predicting, reporting and appointment booking. The functions of this App are simple and practical. Users can get an insight about the current and the predicted customer volume within the working hours of supermarkets/malls remotely. Our App can provide predictions for the upcoming seven days providing users with 3 different alerts: red, yellow and green. Red means the supermarkets/malls have reached their maximum capacity, so users can consider shopping later. Users can also avoid joining a long queue by making an appointment in advance of their visit on our App. By doing this, it not only provides multiple functions to customers, but also improves our accuracy rate of prediction, as we can use both real-time customer volume data (which we assume would come from near-infrared monitors installed in the supermarkets/malls) and booking data from customers. We used Poisson distribution to simulate the number of customers and applied Gaussian processes to make the prediction. We were very excited to find that our model reached the highest average prediction accuracy rate of 95%. To see a demo video of the App, please click here.
This was the first time all team members had attended a Hackathon, so we were all very excited. As a social scientist, I never imagined that I would be able to attend a Hackathon but due to great team work, we were able to deliver a very exciting concept. My team mates, Yifan Li and Yuchen Lu, are data scientists who engaged in research of Bayesian machine learning and implemented the ML prediction system. My role in this project was to clearly figure out the project idea and propose the initial solutions; facilitating brainstorming and group discussions; designing the User Interface and presenting the project. This Hackathon allowed a social scientist like me to act as a product manager and I also applied my interdisciplinary background to solve a real-life problem. This could not have been achieved without the ART-AI MRes year which aims to train PhD students with interdisciplinary backgrounds and enable them to communicate and collaborate with people from multidisciplinary backgrounds.
Despite not being the winner of that track, the judges commented that our App was a great project that demonstrates how tech can be used to enhance people’s life experience after the pandemic.