Besides my main research did some data driven stuff this week. Curve fitting Corona, probability for traffic accidents and thinking about Matlab storage.
I got a pretty fit for the corona spread in China:
It is a logistic function which is exactly what you would expect. At 49 days we are pretty close to saturation. At day 21 it looked for a moment that it was heading towards a logistic function which would top at 45k but alas it climbed much higher. Data available here.
I was looking at the data on traffic related deaths in china. So apparently 18.2 people in 100 000 die from traffic accidents per year. Assume that this data is independent with respect to age, class etc, then the probability that you become 70 and didn't die from a traffic accident is (1- 18.2/(100 000))^70 which gives you about 98.7% to not get run over. Which means that if you know about 54 people from birth that there is about a 50% chance that at least one of them dies in a traffic accidents which seems pretty high to me D: (Last calculation follows from (1- ((1- 18.2/(100 000))70)54 = 0.497 ) )
I needed to store some big matrices for some tests and noticed that I was running out of space which was a bit distressing. I don't really work that much with data so let's have a look at what is going on the storage front. I am using Matlab (2016a) so I expected that it wouldn't be optimised for storage.
