I think my colleagues have said it very well. We have to be careful, obviously, about sharing open data with personal, private information. By and large that does not happen except in specific circumstances.
But there is data that can be shared that would historically or traditionally be considered sensitive that has tremendous value, and I'd like to give you a very quick example of that, particularly around health data and crime data, for example.
Citizens are very interested in crime data. Where is crime occurring, and how frequently is it occurring? You share the data in terms of trending. In statistical data you don't obviously give intimate details around who committed the crime, or who the victim of the crime was.
A real benefit of this happened recently in the U.K. where in one of the cities they analyzed the influx of patients in the city hospitals over a weekend. They combined that with violent crime in the area around bars and restaurants in that city. Between these different agencies—the hospital, the police, and bars and restaurants—they took a collective decision to start serving plastic glasses in bars and restaurants because they realized I think it was up to 40% of the emergency cases in the hospital were caused by fights and violent activities occurring around bars and restaurants.
So here's a great example where you're sharing data in a way that's not sharing personal, private information, it's anonymized data, but it enables a decision to be made that probably wouldn't have happened had that data not been shared in the first place and shared in particular among different government agencies.