Maciamo said:
My extrapolations from the data are based on 2 assumptions:
1) That the case fatality rate is about 0.8% like in South Korea and like the new official estimates mentioned in the BBC article I linked. I think this is a safe assumption.
2) That the diffusion of the virus increase 10x every 8 days in average. That's the actual average I calculated based on real data for each country. However it is true that the number of reported cases depend a lot on the availability of tests, medical staff and the criteria for being allowed to test for Covid-19. Furthermore, my data showed that the diffusion rate was actually slower in Italy and Iran than elsewhere, with a 10x growth achieved after 11 days instead of 8. I checked again with today's data, and it took 12 days to pass from 1700 to 17,000 cases in Italy. If I recalculate based on a 10 fold increase in 12 days instead of 8, the we only get 15 million cases now instead of 150 million. So you are right, not everyone could be infected yet.
A little addendum.
1)
Fatality rate : I said that 0.8-0.9% now seemed a reasonable estimate. But actually, it is very probably going to vary considerably between some countries worldwide depending on the age structure. For example African countries have a very young population. The median age in Africa is barely 19.7 years old! Only a 3% of Africans are over 65 years old in average (see
stats by country) as opposed to 28% in Japan, 23% in Italy, and generally between 18% and 22% in the rest of Europe. Therefore, even if the African whole continent gets infected by Covid-19, the fatality rate is going to be much lower, despite the low availability of ICU units.
The youngest developed countries are Singapore (11%), Israel (13%), then Ireland, Luxembourg, North Macedonia, Albania and Cyprus (all 14% of the population over 65 years old). The United States, Australia and New Zealand (all 16% of the population over 65 years old) are also considerably younger than the European average.
In other words, the CFR should logically be higher in old countries like Italy, Portugal, Germany, Finland, Greece and Japan, but lower in places like the USA and Israel and in all developing countries (including Russia, Mexico, Brazil and China).
2)
Diffusion speed : After reflection, the virus should spread faster in places with high population density and geographic barriers to travel (such as a flat landscape and lots of motorways connecting cities). Based on this, countries like the Netherlands and Belgium are ideal for a fast propagation. That's also why it makes sense that the virus propagates more slowly in Italy and Iran (11-12 days to increased 10 fold, against 6 days in Belgium) as those are both extended and mountainous countries, with two main islands in the case of Italy, and large deserts in Iran. Having had the experience to drive a car through the Apennines in central Italy, I can assure you that this is slow and time consuming. On the other hand, the Po Valley is very flat and very well connected and this is exactly where we observed the fastest spread of the virus. So geography matters. In the USA, remote regions of the Rocky Mountains (Idaho, Montana, Wyoming, Utah) or the Appalachians (West Virginia) are likely to be better sheltered than the big population centres along the coast of around the Great Lakes.
Taking into account all these factors makes the task considering more arduous to forecast the speed and fatality rate in each region. But at least we get a good idea of where it's going to spread fastest at first, and which populations are more at risk age-wise.
Here are maps showing the percentage of the population over 65 in Europe and the USA.
Then, as I mentioned before,
air pollution is an aggravating factor. Northern Italy has no luck as it has one of the worst air pollution combined with one of the oldest population in Europe,
and diffusion is facilitated by a flat landscape and high population density. When we think about it, all the risk factors were present for the region to be hit worst. It's not because it started earlier but because these factors caused the most fatalities the faster within Europe that it attracted all the attention.
The industrial cities of Saxony (Leipzig, Dresden, Zwickau, Chemnitz) in East Germany are also polluted and have one of oldest population in Europe too, so this region is likely to be hit hard as well. Geography slowed the arrival of the virus there, but it will catch up soon.
If I don't need them, they'll come in handy when blowing the leaves or sanding the new patio. 
