As math students, we wanted to apply the knowledge we had acquired so far in helping others understand the spread of a virus better. Particularly, on how effective government measures such as confinement or testing are.
Indeed, faced with a general lack of awareness from the public and an abundance of fake news, we aim to provide a practical and fun tool to inform.
What it does
Our goal is to present data in a pedagogical way, making it as interactive and appealing as possible. Currently lacking the necessary data for specific COVID19 virus simulations, we studied different epidemics models hoping it would give insight into the more specific case of COVID19.
How we built it
Challenges we ran into
The first challenge we ran into was finding mathematical models for an epidemic with explicit or easily approximable solutions. There are models that may have been more representative or that deal with different interesting aspects of the spread: the existence of a vaccine or a non constant population size where we could see the deadliness of a virus, for example. But we were faced with systems of nonlinear differential equations and we do not currently have enough mathematical knowledge to solve them.
Once we had chosen our model, another difficulty we ran into was finding realistic parameter values to represent the spread. Furthermore, since our simulations on MATLAB already took a long time to run for a total population of a million, it wouldn't be optimal to represented a worldwide pandemic.
Accomplishments that we're proud of
It was a really good project to have and even though the SIR and SIS models used here simplified an actual spread, it was a good first approach.We were able to apply concepts we'd learnt in class so far, in particular we used the Galton-Watson process to model our generations of infection spread.
What we learned
What's next for PandemicAnalytics
Since our models simplified reality a lot, we would like to improve it by taking into consideration more parameters and larger scales. We would also like to improve the website interface.
Try It out