I followed a link in this Seth Godin blog post to a video of a presentation given by Tim O'Reilly at a recent conference. The theme of the talk was the importance of data and data accessibility to future web based applications. I did not watch the whole thing, but a point he made about 10 minutes into the talk really struck me. He was talking about smart metering of electricity consumption.
Smart metering has been around for a while and is touted as a way to promote energy conservation by allowing consumers to observe their energy consumption habits. More sophisticated versions would also provide utility companies the information and control needed to allow them to balance loads and control levels of peak demand. In order for utility companies to gain information about consumption of individual appliances it was believed that each appliance was going to need to have some connectivity to the Internet, possibly in the form of unique IP addresses.
The cool thing that O'Reilly said was that people studying data from smart meters have discovered that each type of appliance and even appliances of the same type from different manufacturers have distinct energy consumption profiles. This fact along, with the appropriate reference dataset would allow data miners to identify the type of equipment being used from raw energy consumption data without any connectivity to the Internet or unique identifiers for each piece of equipment.
As O'Reilly himself states in his talk, there are clearly privacy implications for this use of the energy consumption data. However, the more interesting aspect is that it is yet another example of the unanticipated consequences of our ever increasing ability to collect, store and share vast quantities of information.
