Time-Series analyses on the COVID19 Dataset provided by Johns Hopkins University
Our dataset was collected from JHU GitHub. It contains up-to-date information regarding the latest number of COVID-19 confirmed cases and COVID-19 Deaths. Although these models were only ran on domestic (United States) data, international data is also available.
Since the beginnning of the year, COVID-19, has spread across the entire world at an alarming rate. The first case in the United States was recorded in January 2020. As of July 30, 2020 millions of people have been infected, and over 150,000 people in the US, have died from the virus. We wanted to see if we could build a model to see where the numbers are heading.
Our goal is to build a model to project death and confirmation rates on a state and national level in America. Also see if certain actions can be made to decrease the rate of the curve.
- ARMA
- ARIMA
- SARIMAX
The final model selected was SARIMAX model. The RMSE Score for the National Cases Confirmed model was 15,325. The RMSE Score for the National Death Model was 487.
Currently, our models predicted that there will be a continuation of the upward trends in both confirmed cases and death cases, however the amount varied depending on the location.
Based on the analysis performed, it is recommended that cities institute some measure of lockdown protocol. As this has shown promise in mitigating the cases and and keeping the death rate down.