Develops an innovative model for COVID-19, building on previous work. The final SVEIPR model, including modeling of vaccination and asymptomatic infection (“potential” infection, P) is shown to be fairly successful. The report has appropriate references. It leads toward a conclusion justified by the data analysis.
ARMA modeling would be more successful on the log scale.
The residual plot for ARMA shows dramatic heteroskedasticity. When the authors say “We can observe certain stationarity and good fit for this model,” they are failing to see the information in the plot.
Early figures are numbered, but later figures are missing numbers and captions.
Reviewers did not find much to criticise with this project. It only has small novelty relative to previous class projects, but combining a bit of novelty with careful technique meets the requirements of a strong cours project.