Inspiration
Our team was fascinated and inspired by the scientific community's response to the COVID-19 pandemic, especially the work of data scientists in the public health field. Based on this, we decided to create a model to look at what vaccination/immunity rate would be required to stop the spread of a disease. Ultimately, we decided to look at Ebola due to the existing data and the fact that it can fit with a SIRS model.
What it does
Our model simulates the spread of the disease through Masvingo, Zimbabwe using data regarding the population, average family size, and other factors. It is designed to determine the minimum number of vaccines required to limit fatalities and stop the spread through the community, which could hopefully be extrapolated to other areas in West Africa.
How we built it
We adapted a SIRS model for Ebola and included fatality and immunity as options. In addition, we added the number of days for the virus to run its course. Multiproccessing was used to run multiple vaccination rates simultaneously, as otherwise it would have had to run for longer than the hackathon. We stored the number of deaths and infected persons for each vaccination rate after each day in a file and this data was then visualised using MatPlotLib. We created a visulation to display both a zero vaccine and optimal vaccination rate side-by-side.
Challenges we ran into
Unfortunately, we could not determine an optimal vaccination rate other than 100%, and this still led to some fatalities as existing infected persons cannot be vaccinated. The model also took a long time to run, however using multiprocessing we cut that time down to a feasible number, however this still took a while so given more time we may have been able to refine the model to find an optimal vaccination rate.
Accomplishments that we're proud of
We believe our model is a good starting point for the determination of the required vaccine rate, and believe with enough time we could determine it. It allowed us to put into practive our knowledge of modelling and explore how the fields of public health and data science intersect, which we found extremely interesting.
What we learned
We learned about adapting the SIRS model to fit diseases, for example stochastic random updates to simulate the spread throughout the population.
What's next
We will look to refine the model in order to find our required vaccination rate for Ebola. We would then look to see how the model would interact with different communities, and possibly try to model other diseases with human-human spread and existing/possible vaccines.
Log in or sign up for Devpost to join the conversation.