These are the daily data from Italy first spread of Covid-19.
Let's suppose we are on the last day of the lockdown.
Last data, at 18:00 of {{ (inspections[0] || {}).date && inspections[0].date.toDateString() }} was of {{ (inspections[0] || {}).total }} total cases.
Let's try to fit these data in a simpler trend (aka non linear regression on a logistic function).
Prediction tell us that total cases will presumely be around {{ (inspections[0] || {}).max }}.
But removing last data and checking yesterday forecast...
there is a difference of {{ abs((inspections[0] || {}).max - (inspections[1] || {}).max) }} cases
This can be done more and more
The moral
To predict new data based on old ones is not a good practice.
It is necessary to the experts to check the current spreading of the virus, but it is extremely misleading as predicting the future: to use responsibly.
Feel free to explore last days forecasts, (just over on the black line)!
Thanks in advance!
Mauro