Tracking and analyzing your whereabouts with mobile devices
Deborah Estrin, a professor of computer science at UCLA, says the capability of GPS mobile phones to record location over time creates a new wealth of analytics. For example, new opportunities will exist where information is gathered from multiple users that better enables the accuracy of local traffic conditions.
>> Instructor: I want to talk about location over time. So we've been looking at how we can leverage these traces that are basically the time series of our location over time and ways in which they are very telling because of all that can be inferred from them. And they're telling and they're so readily available now obviously because of the fact that we're all carrying around with us mobile phones and this is a technology that is with people and with the pervasiveness that we really haven't seen before; and very importantly it's not just the mobile, everything you hear about here. It's about all of those maps and all of that GIS and all of that information that's behind it. So it's not just that it's real time and real place; it's all of the contacts that it has because it's always connected. And with the full web behind it so it's just a tremendous amount that we can learn and that we can share by recording and analyzing this information; and it's very much a technology as we're looking at it--it's a technology about engagement. So it's not just about passive collection for somebody else. It's about people being engaged with the data that they're collecting. And so this is a little bit different from just the focus complimentary to the focus of your mobile carrier or even Skyhook having data about you or its customers. This is really about individuals having information about themselves and what can they learn from that through the sort of wealth of analytics that we are just hearing about. And yet having started working in this area and being very interested in capturing traces, I do have an interest on the other side of this as well because they're not just telling traces for us as individuals; they're telling traces for us as well. But before you sort of write me off as a over-the-hill privacy Fein, assumed spelling I really want to spend the first part of the talk saying the interesting things that we're doing with them because that is really what has attracted me to this technology is how easy it is to capture these traces and the richness of them in the positive constructive sense of what you can infer and I'll come back at the end to talk about privacy. So consider smart phone applications that are continuously recording your GPS and accelerometer readings and out of that then you have a location activity time series. And for all practical purpose this really gives us a really good sense of what we're doing 24 by 7 again because of all the mapping and model and GIS based information that we can mash it up with. And there are a lot of very interesting aggregating projects that use these traces as a kind of a community mapping tool. My favorite is ways which is using sampling of automotive traffic on uninstrumented streets to build and really create the best source of local and live traffic information when it's used by even a pretty impressively small percentage of the population. But or and there are also applications that don't even require or depend on that minimal level of penetration. And peer is an example of an application that I'm just using my location trace to mash it up with information that the Resources Board maintains about particulate matter concentration. And so then I can see how my commute patterns and my choices about time of day and the paths that I choose to take and as I compare one path to another, I can base on my location activity trace, see my particular pollution exposure levels; or I can compute my particular--do a very personalized carbon calculator. So it's very much an example of what you were seeing before but I'm taking my location trace and mashing it up and sort of doing an integration across the background maps and analytics about the temporal spatial dynamics of air pollution and making it personal. And this is useful, one person at a time so long as that GIS data is back there, I can make use of my location traces to look at how different days or weeks or months or compare one to another. And my favorite example of usage of this was actually last year up in the Bay area. A group of high school kids used the carbon calculator aspect of it to compare--they made an effort across a bunch of classes in a competition who could reduce their carbon impact a class at a time by changing their transportation behaviors. Instead of just using a diary, they ran this location tracing application on their phone and that data is automatically uploaded to their personal peer application and then they designed a little Facebook widget by which they shared their aggregate carbon impact over time.
==== Transcribed by Automatic Sync Technologies ====