Predictive surveillance uses audience segmentation and behavioural data to model future behaviour based on historical patterns.

Hmm, I need to write an 'how predictive surveillance works' article don't I? Should how existing audience segmentation is already most of the way there...

Hmm, I really need to write an article explaining how predictive surveillance likely works (behavioural prediction). Audience analytics → behavioural segmentation → propensity modelling → trajectory prediction → persistent identity → predictive surveillance Audience segmentation systems (like my toy one https://lnkd.in/e995AzE4) *already* have the data for this. It's establishing groups (segmentation), then you can use characterisations of behaviours (centroids) and how those tend to move (acceleration / direction vectors) *for similar members in the group*. By observing enough signals over enough time for enough people, yes, these can become predictive of future behaviour - not prescient, just statistically modelling how the behaviour moves across time. It's basically how advertising works out what ad to show / how Facebook or Twitter build groups (hate, subject topic, purchase behaviour etc) to 'nudge' behaviours to something they want you to do. Terribly unethical, terribly complex, so nobody really understands *they already