Yeah I work with this stuff and machine learning data is everything these days. The SIDs thing was sick though - statistics work with things that are likely to be true but it doesn’t mean it is absolutely true. It’s generalisation versus manipulation of data to make it look true. I personally love this stuff, I look for the real truth in data.
I predicted 7 days ago we would have 400 deaths in the UK on Wednesday and got that perfect because I understood that COVID affects 70% fewer of those hit by delta, so the data meant that we would only see a rise of 33% from the first peak as the death data peaks 2 weeks after that of the new cases! The big rise in data looked huge but it was obvious to me that we didn’t have much to panic about, regardless, as I looked at the big rise in new cases as 70% lower per day.
Really well explained by the way