Sai Srikanth Lakkimsetty, Sayan Biswas, Derrie Susan Varghese, Sneha Agarwal
We have used the SIR model to fit the transmission of the Novel Coronavirus SARS-2 (Coronavirus) deadly disease to real world data. Additionally, we also wanted to figure out a way to explain how uncertain we were about that model being right.
This model puts everyone in one of the three categories: Susceptible, Infected or Resistant. The model implements the ordinary differential equations (ODEs) that govern the respective populations and infer the posterior distributions using an inference algorithm (MCMC). Additionaly, we forecast the S-I-R populations for the next 90 days. We implement the concept of interventions in simulation modeling context and forecast those trajectories. Finally, we answer a few counterfactual questions by leveraging the ability to implement interventions.
See video abstract for COVID-19 SIR Model
- notebooks/SIR-model.ipyb: the code for the entire project is available in this file. It is entirely reproducible.
- img: this folder contains all the images used in the notebook file.
- slides: this folder contains the slides which provides a brief overview of this project.
- Data: John Hopkins - Whiting School of Engineering: CSSEGISandData
- Data: European Centre for Disease Prevention and Control: COVID-19 Data
- Predicting coronavirus cases