Web 2.0 Summit 2009 Discussion: Humans as sensors Markus Tripp: - Mobilizy Gmbh: established 2009 - Augmented reality world browser (wikitude), launched on Oct. 2008, - Wikitude.me: community created content (the social web interface to Wikitude) similar to google earth- everyone can create content - Mobile phone has GPS and compass and camera. We overlay information on the background seen on the camera relevant on the surface. --------------- Sharon Biggar: Path Intelligence, two products: FootPath and QPath - online world is good at collection data on user experience; what products I'm looking at, purchasing or not purchasing - analogous to google analytics (in the online world). Instead of detecting clicks, we're detecting anonymous pings from cell phones, does not require an app to be downloaded onto your phone, we are installing sensors into high traffic areas and then we passively detect anonymous signals that broadcast by your cell phone. By understanding what crowd is doing, we understand what people are interested in "right now". It tends leads any other information that we get- even tweets. Help improve businesses what people are interested in now. We have a technology which locates the phones highly accurately in indoors. - detecting real time interest of figuring out what people are interested in now - i.e. bring the same type of online data collection to off-line retailer industry - Path intelligence takes online innovations and bring it to the real world by installing sensors and correctly locating mobile phone indoors, using crowd data to help businesses improve in the real world in real time. --------------- Di-Ann Eisnor: Platial, Waze is free crowd-sourced navigation and real-time traffic. Started in Israel and well past critical mass, Waze now has top quality map, navigation and traffic data in the country. Real time crowd sourcing + live mapping. - crowd sourcing: twitter feeds to standard map information - waze: crowd source meets real-time information- creating waze. Waze is a crowd source navigation and real-life traffic information. Just like any other navigation application, it gives you a route, turn-by-turn instruction, it's free. The difference is that all the information is coming from crowd. - turn on the phone and drive, it's passive. We're taking information on speed, direction of the roads- taking tiger data, display map from the census burau and turn it into the navigatable map together with our users. On the traffic side, we're able to make inferences based on historic speeds, able to tell whether you are driving faster or slower than what we have in the database for a particular road and we are able to draw traffic based on this kind of information. This data is delivered on-the-fly, 30 seconds delay. The map data is updated every single day. It launched at Israel at the beginning of 2009 and reached critical mass in 6 months. We have superior map and real time traffic information for the country, 20% smartphone users (220K drivers) using the application. Transactional cartography- crowd sourcing merging with real time data. We can re-route you based on what we see in the world. We used to map objects, then people and their migration patterns and now mapping actual processes Couple academics called this "weather mapping for social processes in the 21st century". --------------- Debra Estrin Allow every day people to not just provide data to important things such as traffic but also to author their own data compaigns, for issues important to them or to their community. Smart phones and web services provide phemonally scalable "make a case technology" and they do this through explicitly captured information or automatically captured information. Let me give you examples to both of these, the examples clearly help. Explicit data capture example: What's invasive! national park service- it's live. Park service employees and/or visitor can download the app to their smartphones and use it to tag the locations of the invasive plant species threatening the eco system. You can participate in the site by providing data for your favorite wilderness area on your hike area. Upload the top ten most wanted weeds for your neighborhood that may not get represented with a site. The general technology of the participatory sensing is a story of mobile to web and making it very easy to create your own data compaigns as well as contribute other people's. Implicit data capture example: Biketastic. Information can be automatically capture data from other sensors on the phone such as location and motion. Tool for cycle commuters to create and share information about bicycle commuting. Automatically captured information, its an app you start and place it on your bike. The information from GPS and accelerometer is captured, activity classified and uploaded and displayed on the map. You get information on duration of the ride, start and stop, how long you spend on noisy, polluted, dangerous traffic intersections. How smooth is the ride or not. The same technology contributing to waze's maps, the same sources can contribute to the information you want to contribute for your community. Traces are powerful tool for both personal and community behaviour change. Prius dash board - refer to as a model for providing real time feedback to create behaviour change to be able to see your measured behaviour. We can use our location traces like the ones bicycle commuters are using. We can use our location traces to understand the implications of where we spend our time and what we're spend our time doing. What is the aggregated air pollution that I'm exposed on days when I commute during heavy traffic hours versus when I commute at off hours. In health care applications, same type of personal traces when I do care plan adaptation. One of the nice things about these applications is that they can scale down, not only up. You don't need high market share or penetration to gain the benefit of these applications. You have to have a value to the individual whose collecting their data and using it to understand something about themselves or about their environment. Q&A session: Q: What types data made available by the smartphones which was not captured before? Di-Ann Eisor - geospace: acceleration, speed, direction becomes ubiqutious, opens tremendous amount of possibility - being able to go from specific nodes or humans as sensors and being able to aggregate the information up so that we can see full processes and full systems was not really available with real-time analysis processing and visulization we're seeing today Sharron Biggar Having access data in 365 days makes a huge data. Being permanently installing and permanently watching what's happening allows us to witness rare events that happens so often, emergency events are good example. For instance, there was a fire in one of sites, we now know how people escaped from that building in real-life emergency situation. That information was given to emergency services, architects, and engineering to help design buildings that react to how people really react to the emergency situations. Markus Tripp With this application, since the mobile phone display is small, we cannot display all the information. Use the sensors to filter the relevant information for the user. For instance, I'm not interested in seeing cheap hotels since I'm leaving there. Q: Some applications pick up data explicitly and some implicity. So, how do we trust the data (implicit and explicit)? Debra Estrin For both implicit and explicit data, giving people visibility back into the data they've contributed both through automated capture as well as through explicit. Also, you do have eyes in the process- there is still human in the loop in terms of causing alerts, acting on the data. These systems are not entirely automated (humans to sensors to automated response). Someone is there sort of watching and we can put in semi-automated tools that notice the difference between harding grass and starfus which looks similar, you can do image recognition. Start introducing back-end tools to try to find those things. Huge part of it is visibility- legible feedback to people whether its the operator or individual on what data is actually being captured. Q: What types of critical mass of people you need in an area to drive in area Di-Ann Eisor It depends on the application and business model. For them, Israel, incubator testbed, 0.5% of metro considered to be critical mass. Debra Estrin It is great if you can go out with an application that has a scale down property- you don't have actually any critical mass, you are starting to provide value, as you get more, you get the aggregate value. Think about search. The search did not start out by having the value of noticing what we search for. It just provided a value that started getting people searching. A huge market of seeing of what we searched grew out of it. If we build our apps and market that grew up that way naturally, we see that stuff happening faster. Q: How do you plan to make money with your apps? Sharron Biggar Business model which brings the sensors into the real world. The malls, amusement parks are paying them to provide information- want to know how people are behaving.