The UK newspapers are currently covering the tragic case of a man who died after “contact” with a member of the UK police force during demonstrations at the G20 summit in April this year. Interestingly, the S+ folks have a customer in a US police force that decided to use S+ predictive analytics to try and prevent similar cases by analysing police data (i.e. “policing events”). They looked for possible patterns that could indicate where a police officer may need further training or reassignment to avoid “negative situations” occuring. This project was considered a success, although naturally is not one that will ever be a public reference …
Obviously the same idea can be applied to any high-risk business involving skilled and trained personnel who are in a position to cause harm to the organization, their customers, or themselves. Consider the possible detection of rogue traders that bring down investment banks, based on their previous trades and market positions. This means identifying useful data, capturing the necessary events, and running the analytic models against them. In CEP this would truly be “sense and respond”…
