Suddenly it is all so clear: The next big corporate race is going to be getting the biggest mobile crowdsourcing network.
It’s been obvious for a little while that mobile location-aware communications devices like the Nokia N95 will be ubiquitous five years from now — they’re the Star Trek Tricorder meme made real (though the UI will look more like the iPhone’s, I hope). What wasn’t so obvious (at least to me, until this week) is that these devices will make great georeferenced information collection networks, perfect for data-mining. The corporate challenge: Building, buying or otherwise gaining access to such a network.
The evidence (and a blind alley):
There have been two acquisitions that have been widely reported in the GIS-o-sphere this past week: GPS device maker TomTom’s planned acquisition of digital mapping company Tele Atlas and Google’s acquisition of aerial imagery provider ImageAmerica.
(Google’s purchase has generated a lot of press but is the least relevant to this story: The company has been renting the services of aerial digital imagery providers (in the US and also in Australia) and I suspect it is just cheaper to own the planes and cameras if you plan on updating the most viewed regions more often. It’s a logical next step in Google’s mapping mission: it increases their ability to generate new imagery quickly without having to depend on a third party to allocate resources as they see fit. (That, and Google Earth CTO Michael Jones loves gigapixel cameras:-) )
The TomTom-Tele Atlas merger presages the coming of such geospatial collection networks, and newly minted Googler Ed Parsons sees its implications right away:
Without community generated content, in a online future it will not be possible to provide the expected level of currency of data – Strong stuff but hard to argue with.
Daily Wireless excerpts a Newsweek article that struck me: Both Tele Atlas and competitor NAVTEQ already employ hundreds of “road warriors” in the field to gather data that satellites can’t catch: traffic signs, one-way streets, points of interest…
“At the end of the day, there’s no substitute for going out there and capturing the real world ourselves,” says Navteq CEO Judson Green.
How to scale that? Build a crowdsourcing network, so that others can capture the real world for you — using devices that companies like TomTom (and Garmin, and Nokia) can build.
Geospatial crowdsourcing already has a impressive working example: OpenStreetMap.org, where you can (manually) upload your own GPS tracks to create a collective world road map or edit existing data. And now it turns out that Google has been piloting something similar in India, as explained by Michael Jones last week at a conference. Dan Karran has transcribed the relevant bit of Michael Jones’s talk:
We have a pilot program running in India. We’ve done about 50 cities now, in their completeness, with driving directions and everything – completely done by having locals use some software we haven’t released publicly to draw their city on top of our photo imagery.
From static to live
So far we’ve only been looking at collecting data with a long half life, such as the locations of gas stations and roundabouts. But when there are enough human “sensors” in the network, Michael Jones explains,
[…] it has the advantage of, when the road is closed, you can click on that road and say it’s closed today. If you’re having a block party, you can say the block is closed this day. Traffic data that’s up to date every day.”
Google’s solution doesn’t (yet) involve the use of a mobile gadget to contribute live data, but such solutions already exist elsewhere. Already in 2006, new Honda Civics driving on Japan’s roads could wirelessly upload live telematics data to a central server to calculate road congestion. For a fee, you can get that information delivered to your car dashboard.
Getting location-aware mobile devices to contribute data to such networks means less work for humans, as there is no need to manually georeference the data. The easiest milestone on this road (pardon the pun) is live traffic reporting, as it doesn’t require the active participation of a human beyond driving. In the future, if TomsTom get the ability to transmit live telemetrics on busy highways via wireless data networks, you might start saving lives if a traffic accident in dense fog involving a TomTom-equipped car is reported in time to those following soon after.
Traffic monitoring may be a special case — cars are nearly always in the line of sight of GPS devices, whereas humans are more often than not inside, or in a concrete city canyon. So how to get humans to actively contribute to a georeferenced crowdsourcing network, other than by paying them, which isn’t scalable? You could appeal to their nobler selves (the open source model), you could make it fun (Google Earth Community), or you could appeal to baser motives — let the contributor get something out of it, such as access to the collective wisdom of the network (for example, to find the rooms with the best views in hotels as reported by previous visitors). But above all, it should be easy for the contributor.
This is where this week’s under-reported acquisition plays a part: Nokia bought Twango. Twango is a media sharing web application for mobile devices. Unlike its main competitor Shozu, which offers to forward your media to your web sharing app of choice (such as YouTube, Flickr and others), Twango wants you to store all your media in one place, on its servers, and share it using its URLs as a destination. With Twango soon installed by default on new Nokias mobiles, Nokia will be making a play to capture the generated media of its hundreds of millions of mobile phone users, so that network effects are its to dispense — with uploaded videos and photos providing eyeballs to advertisers.
Shozu gives users more choice when it comes to which network they want to contribute to, and I hope that this is the model that catches on, as it is in the users’ best interests to have device agnosticism when it comes to networks, and network agnosticism when it comes to devices. Meanwhile, Shozu already georeferences images for you if you have a GPS-enabled phone, though presumably Twango will soon follow — at the moment you have to tag for places manually.
So: Car navigation devices are going to communicate data in both directions, live; mobile phone manufacturers are going to load their phones with default software in a bid to capture and host the generated media; and other network web apps are going make sure they have APIs to which these devices can post, perhaps even automatically. With Nokia’s GPS phones already providing a fee-based navigation service, it’s inevitable that all these manufacturers are converging on devices with similar functionality, while the information gathered from these gadgets is going to be valuable to whoever gets access to them.
4 thoughts on “Coming up next: The human sensor web”
It’s somewhat cracked me up, in a sense, over the last few years — that the industry has seemed hesitant in understanding where they would stand in the coming years in light of initiatives such as Google Earth or other mapping portals/applications. But the reasons I chuckle, is simply because I knew that it would be a matter of time before anyone realized that regardless of the application — the GIS and data aspects are the primary focus of what will drive the industry moving forward. And in so many ways, it’s been baffling to see so many having missed the boat, so to speak, concentrating more on areas that really didn’t make much sense to compete in — when in fact they had the ability to charge ahead all along, leveraging their core strengths.
It’s all about, and it will always be primarily about — the data. Without the data, you have no GIS. Without the data, you have no application.
It’s been an interesting observation, however. I feel that the developments have been a good part of accelerating the field — and generating all kinds of competition, helping to substantially gain a momentum that otherwise wasn’t present even a couple years ago.
As in gears, it just takes a matter of time before both teeth click together to take us all to the next step.
I still view the greater industry (or community) as being anything but unified. That persists as the major issue overall moving forward. But in time, perhaps those teeth will click as well.
thanks for the great blog. quite interesting content on mapping.
Note that well before 2006, many of japan’s roads had metal coils embedded in them, sensing vehicle frequency.. so they essentially have the ultimate traffic information system, which you can have piped to any ordinary car gps system for few dollars per year. Even in 2001 or so.. i recall seeing the roads colored with red and green on many car’s navigation units.
In the US, I belive http://www.dash.net/ will bring the concept of “wirelessly upload live telematics data” so anyone with a dash device can benefit.
AI, you have an interesting point about the Japanese sensor systems. I’ve read of those as well.
This is perhaps off-topic, but I’d like to add a note about the potentials of such systems. (Though I hardly see us cowboy Americans ever wanting this.)
With such a sensor system, or if more elaborate — it’s conceivable to think that law enforcement would then be able to effectively force a vehicle to pull-over via a communications command to the vehicle via GPS. This would reduce or eliminate any opportunity for a fleeing motorist scenario and dramatically reduce risk of a high-speed chase which places everyone, including those involved in harm’s way. (But primarily the innocents who could get tangled up in the potential crashes that could occur.)
Also, it’d be nice if law enforcement were equipped with vectorizing license plate readers, wouldn’t it? Why increase the potential for a traffic chase or any amount of road-side conflicts, if a motorist is passing by, speeding? Send them a ticket in the mail.
Besides, with such a system, how difficult would it be then to run that information through the database — identify the vehicle as belonging to a person wanted of any range of legal issues — if detected, shut the vehicle down, turn-around and deal with the subject.
Not too hard, is it? Everyone’s safer. The risk potential has become highly diminished, and police officers can no longer use traffic violations as excuses, which also fills-up our courts with endless unneeded legal cases that do nothing but serve to flood the courts and generate inefficiencies. (Not to mention, the liklihood of accusations of violations, which most municipalities have learned generates them a very hefty bottom-line in revenues every year.)
What I understand so far from this article about “crowdsourcing networks” (nice term!) is that there are only a few known business models and probably less which make money:
– traffic information (also possible through mobile networks)
– enhanced base maps (landmarks)
– location based advertisement
– increased speed of collection of marketing intelligence
These business models exist already, and crowdsourcing networks will increase their efficiency quite substantially. So this is the expected progress.
On the other hand since neogeography is great fun, I wonder which other business models will come up. In the marketing field I have not seen so far (I love to be corrected) a very deep impact of geomarketing, GIS or neogeography, only a partial one. Most of the information processed and aggregated for decision making rarely does not have any meaningful geo-attributes. So I am wondering what increased geo-information availability (with partly poor geocoding) will result into?
Maybe I should look more often into the “NeoGeo Jobs” section?
Comments are closed.