basingwerk wrote:
I was thinking that any data can be put in a data warehouse, and people will find innovative uses for it.
Quote:
They won’t just collect the wrong data or the right data. They’ll collect ALL the data that can be economically collected, and then find new things to do with it.
Yes, they
will find innovative uses for it, a fact which anyone with a modicum of common sense should find deeply disturbing.
'Data dredges', as they are known, find 'statistically significant' links between unrelated sets of data, and, taken out of context, lead to all sorts of dodgy conclusions.
These 'data dredges' have been responsible for all sorts of shenanigans - like the banning of innocuous household chemicals which have been used for centuries.
Quote:
That will only collect data that can be got at when a crash occurs, which is diametrically opposite SafeSpeed’s view that near-misses are important too. As you say in your .sig, when two opposite points of view are expressed with equal intensity, the truth is not in the middle!
No, I wasn't opposing anything. Crashes are easy to detect, but how is a system going to reliably detect near misses?
It's all very well saying that one should therefore record all data, but you should know 1) how much bandwith that would require, especially with the vital video data, and 2) you'd end up with tens or hundreds of thousands of hours worth of data for each incident - how are you going to sift through that in any meaningful way?
Cheers
Peter