I think time span which is available now to look at is very short to consider it time-series. Moreover, its a single breakout which is spreading not breakout at different point of time. So what we have is just one unit of time to look at.
Hi Mohammed, answering to this question is a little hard. It depends on the data and selected variables. Generally we fits various types of models then by comparing their accuracy we can find the best one.
I did an analysis using multivariate time series. But I'm worried about the accuracy. I think better to use encoded daily reports of many countries. Then we can have a better dataset for a simple prediction. My suggestion is to do a normal multivariate prediction without considering time.