Dryce
Hardcore MB Enthusiast
- Joined
- May 17, 2006
- Messages
- 10,911
- Car
- ..
The insurance companies spend a lot of money researching activities like this.
By looking at people's lifestyles, their social media activities, health, possibly even holiday destinations, what you watch on tv, spending habits and a load of other things you get up to it enables them to paint a pretty good picture of each individual.
It's not foolproof but pretty good nevertheless.
It's probably pretty rubbish. Why?
Well as with a lot of data analysis you can only analyse what you can see. Trying to correlate this with the real world - particularly the people you can't see may mean that it's not meaningful.
A few people publish a lot about their lives on social media. Some people publish a bit. A lot of people don't actively publish. Different people people congregate in different parts of the social media web. A lot of people are anonymous.
So you can do great demos with social media stuff to demonstrate the value of it. But iot's actually hard to make things work beyond that. So you can get managemennt excited but a year or two down the line it may be hard to show that it was really all it was cracked up to be.
I would be suspicious that this isn't about accessing data to assess risk - it's about mining data to discover stuff they don't know and try and apply it to risk - and also mining it for marketing purposes.