I have prepared a time series model using FB Prophet for making forecasts. The model forecasts for the coming 30 days and my data ranges from Jan 2019 until Mar 2020 both months inclusive with all the dates filled in. The model has been built specifically for the UK market
I have already taken care of the following:
My question is, that how do I take care of the current COVID-19 situation into the same model? The cases that I am trying to forecast are also dependent on the previous data at least from Jan 2020. So in order to forecast I need to take into account the current coronavirus situation as well that would impact my forecasts apart from seasonality and holidaying effect.
How should I achieve this?
Haven't tried it but one thing I have in mind is to add covid-19 as a one-time holiday. You can see an example for adding a one-time promo as a pseudo-holiday in the following link: https://towardsdatascience.com/forecasting-in-python-with-facebook-prophet-29810eb57e66
I assume that it can be applied to a negative effect, caudes by covid-19,