Summary: 

The author ran an ARIMA model to make a prediction about how the
chikungunya virus will progress in certain Caribbean countries. The
epidemic has lasted a little under two years, and there have been over
60 thousand cases in the Americas. The author has counts per week as
reported by national agencies (?). She adjusts for certain errors in
the data entry process, which were detected by a jump at week 72.

She ran a SIR compartment model. It is a little unclear what the
results are for this section. It is clear that the model is not the
best one possible.

She ran a multi-country SIR model for St. Martin and
St. Barthelemy. This accounted for the transfer of the disease from
one country to the other through people who arrive on planes.

She ran an ARIMA model for forecasting and estimated the future counts
per country.



Dear Purvasha, 

I read section 4 of your paper. Your report is very well written, and
I understood your methods and the problems with each one.

Your exploratory data analysis is very clear and interesting.

It might be interesting to draw a sketch of the compartment mode as
boxes. That might make it clearer to the reader what the stages are.

You might want to include a section at the beginning of your results
(section 4.0 or something like that) that lists the different models
used. You probably had this in an earlier methods section, but it
might be helpful to introduce the results so it becomes
self-contained. That way if someone is skimming your paper they can
get a good idea of your results from the first paragraph in section 4.

See the comments on the PDF.