GigaOM » geolocation

Now there’s an app to help you dodge bullets

Amanda Alvarez — Fri, 26 Apr 2013 21:26:05 +0000

Some researchers from Vanderbilt have developed a new app and hardware module that will help you find the direction of gunfire. The research team used the sonic signatures associated with firing to pinpoint its location, and put this on an Android smartphone map.

Originally developed for the Department of Defense, acoustic shockwave bearing estimation was designed to help soldiers locate snipers. The technology takes advantage of the properties associated with gunfire – the initial flash of the muzzle blast and the shockwaves that follow. The supersonic speeds and whizzes of bullets can be tracked with microphones and a really precise clock hooked up to a microprocessor. These sensor nodes communicate with smartphones via Bluetooth; data from a few differently placed sensor nodes are required to triangulate the location of the gunshots.

The sniper location system was built into combat helmets, but the research team has now updated it for smartphones with funding from DARPA. Some nodes are still required, so civilian use may not be practical. But the researchers think security details or police squad cars could make use of the smartphone version.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Why the collision of big data and privacy will require a new realpolitik

David Meyer — Mon, 25 Mar 2013 17:36:44 +0000

When it comes to protecting privacy in the digital age, anonymization is a terrifically important concept. In the context of the location data collected by so many mobile apps these days, it generally refers to the decoupling of the location data from identifiers such as the user’s name or phone number. Used in this way, anonymization is supposed to allow the collection of huge amounts of information for business purposes while minimizing the risks if, for example, someone were to hack the developer’s database.

Except, according to research published in Scientific Reports on Monday, people’s day-to-day movement is usually so predictable that even anonymized location data can be linked to individuals with relative ease if correlated with a piece of outside information. Why? Because our movement patterns give us away.

The paper, entitled Unique in the Crowd: The privacy bounds of human mobility, took an anonymized dataset from an unidentified mobile operator containing call information for around 1.5 million users over 14 months. The purpose of the study was to figure out how many data points — based on time and location — were needed to identify individual users. The answer, for 95 percent of the “anonymous” users logged in that database, was just four.

From the paper:

“We showed that the uniqueness of human mobility traces is high, thereby emphasizing the importance of the idiosyncrasy of human movements for individual privacy. Indeed, this uniqueness means that little outside information is needed to re-identify the trace of a targeted individual even in a sparse, large-scale, and coarse mobility dataset. Given the amount of information that can be inferred from mobility data, as well as the potentially large number of simply anonymized mobility datasets available, this is a growing concern.”

Just because you’re paranoid…

For those already worrying about the privacy-busting implications of mobile device use, this should come as no surprise. As CIA CTO Ira “Gus” Hunt stressed last week at GigaOM’s Structure:Data conference, mobility and security do not go hand-in-hand. You can be constantly tracked through your mobile device, even when it is switched off. What’s more, those sensors you’re pairing with your device make it ridiculously easy to identify you.

From Hunt’s speech:

“You guys know the Fitbit, right? It’s just a simple three-axis accelerometer. We like these things because they don’t have any – well, I won’t go into that [laughter]. What happens is, they discovered that just simply by looking at the data what they can find out is with pretty good accuracy what your gender is, whether you’re tall or you’re short, whether you’re heavy or light, but what’s really most intriguing is that you can be 100 percent guaranteed to be identified by simply your gait – how you walk.”

One of the explicit purposes of Unique in the Crowd was to raise awareness. As the authors put it: “these findings represent fundamental constraints to an individual’s privacy and have important implications for the design of frameworks and institutions dedicated to protect the privacy of individuals.”

But this isn’t just about mobility; it’s also about the implications of our big data society. These are effectively two sides of the same coin – mobile devices make it easy to collect data, while big data capabilities make it increasingly trivial to take the resulting mass of supposedly anonymized data and tease out the kind of specificity that the anonymizers were trying to erase.

This was precisely the sort of problem foreseen by Europe’s cybersecurity agency, ENISA, a few months back when evaluating the continent’s proposed “right to be forgotten”. If a citizen really wants all traces of their personal data removed from the web, ENISA pointed out, that would have to mean removing their data from anonymized datasets as well as from more obvious repositories such as social networks and search indices.

As ENISA said at the time:

“Removing forgotten information from all aggregated or derived forms may present a significant technical challenge. On the other hand, not removing such information from aggregated forms is risky, because it may be possible to infer the forgotten raw information by correlating different aggregated forms.”

Shall we just give up now?

The Unique in the Crowd authors stressed in a BBC interview that “we really don’t think that we should stop collecting or using this data — there’s way too much to gain for all of us — companies, scientists, and users.” So what can be done?

Personally speaking, I have been writing about issues around data privacy for many years, and I still cannot see any easy solution to this problem. If it were simply a case of which side of the argument carries more weight, I would have no hesitation in siding with the privacy brigade: selling data to advertisers in order to fund that “free” app does not justify the creation of a surveillance society.

But it’s just not that simple. That Fitbit is also trying to help you keep fit — the fact that it can identify you by accident doesn’t change that fact. Mobile operators’ datasets help keep their networks running. Location-based services don’t work without location. We even hope big data capabilities will help us fight diseases and socio-economics problems. And, most importantly, despite the fact that most people in the U.S. and European Union insist they want better data privacy, we see time and again that this desire doesn’t translate into action – people still give up their data without much consideration.

What we need is a new realpolitik for data privacy. We are not going to stop all this data collection, so we need to develop workable guidelines for protecting people. Those developing data-centric products also have to start thinking responsibly – and so do the privacy brigade. Neither camp will entirely get its way: there will be greater regulation of data privacy, one way or another, but the masses will also not be rising up against the data barons anytime soon.

There needs to be better regulation that works in practice – unlike Europe’s messy cookie law or the “right to be forgotten”. It may be that the restrictions will need to be on the use of data rather than its collection, as proposed in a recent World Economic Forum report. However, regulators tend not to be very proactive, particularly when the risks, while inevitable, remain mostly theoretical.

I suspect the really useful regulation will come some way down the line, as a reactive measure. I just shudder to think what event will necessitate it.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

ESRI buys location services provider Geoloqi

Ryan Kim — Mon, 15 Oct 2012 21:34:08 +0000

Geoloqi, a new platform for real-time location services,launched in February and caught the attention of mobile developers who liked the way it allowed them to easily and efficiently add background location tracking to their apps. The Portland, Ore. company, founded by Amber Case, also turned the head of ESRI, a longtime provider of geographic information systems and geo-planning tools.

ESRI announced today that it has bought Geoloqi, which will anchor a new Esri Research and Development Center in Portland. The purchase price was not disclosed. Case will become the director of the new R&D center.

And they’ve already got a jointly developed product: a new geo-coding enhancement for Geoloqi’s API, so developers will be able to trigger actions in their apps based on real-world addresses and locations, not just latitude and longitude data. Geoloqi and ESRI are also offering a mapping library that can be an alternative to Apple’s default library for iOS developers.

The deal came about after Case met with ESRI and talked about possible investments or partnerships. Eventually, it looked like the best scenario was a sale, Case told me in an interview. She said Geoloqi, which will continue to offer developer tools, will be able to sell to a much wider audience through ESRI, which has some 300,000 customers.

ESRI will be able to supplement its vast mapping and points of information data in its ArcGIS product with real-time location data drawn from Geoloqi. And it will be able to sell its products to mobile developers, who may have been intimidated by ESRI’s earlier products. ESRI has been around since 1969 and sells its products to national governments, Fortune 500 companies and thousands of colleges and universities.

“Many mobile developers didn’t know about ESRI. It was written for GIS developers. We’re making it accessible to all developers. Our objective is to wrap up our tools and API so a whole new industry can use these tools,” Case told me in an interview.

It’s a quick exit for Geoloqi, which was founded in 2010 and has raised $350,000 to date. It recently teamed with Appcelerator to make a Geoloqi module available for Titanium developers. Case said she expects to add another 20 people to the Portland center in the next two years.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

How better location data could mean more-targeted mobile ads

Derrick Harris — Wed, 19 Sep 2012 18:30:46 +0000

The pipe dream for companies trying to make money from mobile data is to know exactly when users are near a particular business and to hit them with the perfect ads to win their business before they’ve moved on. As it stands, however, it’s still quite difficult for software to determine what businesses consumers actually frequent, as opposed to the ones they just happen to hang out by. However, adding the right data into the mix or taking mobile payments mainstream could change everything.

Prospects vs. shoppers

The gap between where we are and where we want to be struck me on Monday as I was reading about Sense Networks’ new Retail Retargeting service. The service aims to help mobile advertisers reach consumers based on where they’ve been and what businesses they’re likely to be nearby again. So, if someone is nearby a Target three times a day, five days a week, that might be someone Target or one of its competitors wants to, well, target with an ad before their next trip.

However, as I looked at Sense’s list of the “top 50 brands with the biggest retail retargeting opportunity in mobile,” I noticed a problem — although I’m almost always within the presence of one of them, I only frequent a few of them. While I always seem to find myself nearby a Subway (ranked highly on Sense’s list because of its omnipresent nature, presumably), I can’t imagine the company could place an ad on Angry Birds good enough to lure me inside.

Source: Sense Networks

Kevin Hannan, vice president of product management at Sense, told me that although its targeting isn’t yet perfect, it does make a difference for advertisers. Although the goal is to identify “shoppers” (i.e., people who actually shop at businesses) — something Sense’s machine-learning algorithms can help it discern when they gather enough info on someone — there’s certainly value just being able to identify “prospects” (i.e., people in close enough proximity to a business that they might become shoppers).

Here’s how David Shim, founder and CEO of location analytics startup Placed Analytics, whose product targets mobile developers rather than advertisers, described the various levels of accuracy in an email to me:

With monetization, content owners are interested in a wider list of places users are nearby to increase the pool of potential advertisers. In terms of competitive intelligence, businesses are interested in fine grained data that can assign a person to a place, and furthermore, layer in index, affinity, and path. Location analytics from a product use case requires a combination of precision and coverage to what places could the user have gone to, and what placed did they ultimately choose to visit.

Building smarter algorithms

And it’s not as if geo-location analysis can’t be done more accurately. Whereas it can be difficult to decipher a specific location in densely populated urban areas, it’s a lot easier to determine when someone is inside a Walmart or other store that takes up a large chunk of land.

Algorithms can be tuned to account for the probabilities that a particular user is any of a given number of businesses, too. “[A] person standing in front of a 20-story building could potentially be at any number of businesses in that building, but realistically, there is a low probability that this person is at the law firm on the 19th floor, the hedge fund on the 17th, or the call center of the 7th,” Shim wrote. “Rather it is more likely that the user is at the Starbucks, McDonald’s, and Bank of America on the first floor.”

A report from Placed indicating when users are usually nearby.

Other data, such as knowing when businesses are open and closed — something Placed’s business database includes — can help narrow the possibility even more by eliminating locations that aren’t open. Saga, a personal-assistant app for the iPhone, will naturally assume you’re at the most-popular of a group of possible places, but learns users’ routines as they tell the app where they actually are. Although Hannan said Sense is not currently looking at adding data from external sources, I think it could be an effective method for improving accuracy in urban areas.

Teaching Saga where you are.

Knowing how often Saga, Foursquare or Facebook users check into certain locations — even anonymously and in aggregate — could help algorithms make better judgments about where someone likely is. As Foursquare founder Dennis Crowley is slated to discuss Thursday at our Mobilize conference in San Francisco, companies like his are sitting on a treasure trove of data about where consumers spend their time in cities across the country. Based on correlations to other behavior (e.g., people who frequent Starbucks also frequent book stores) or just on the sheer popularity of a given place, geo-location models could give a more meaningful weight to each possible location.

Mobile payments to the rescue?

Rather, Hannan seems to be banking on the advent of mobile payments to drive the value of location analytics. Accuracy aside, he thinks a major limiter to the effectiveness of services like Sense’s is the difficulty of measuring conversions, or when a click on an ad translates into a purchase. There just isn’t an easy way to tell when someone clicking on a mobile ad actually goes into a store and buys something as a result.

An easy way to alleviate this problem, Hannan said, would be for more businesses to accept coupons displayed on mobile devices, but a better way might just be for more commerce to carry out on the devices themselves. The trick, Hannan said, will be making the process as frictionless as possible for consumers. One-click purchasing, for example, even if it still requires going to pick something up at the store (like PayPal is now testing out), means that companies will get a much-clearer picture of how ad clicks convert into actual purchases.

However, he cautioned, “If i have to enter my credit card number, this isn’t going to work.”

But all moves toward making location data more valuable to advertisters and, in theory, consumers are good news at this point. “It’s going to take some time for us and other players in the industry to show this is worth the time and is an effective way of targeting advertising,” Hannan said.

The silver lining is that getting in during the early stages of the learning curve of location-based targeting isn’t too costly. The models might not yet be accurate to a dime and mobile payments are in their nascency, Hannan explained, but mobile real estate costs so much less than web real estate that it doesn’t take too much improvement to deliver a return on the mobile investment. Already, he said, using Sense’s full complement of services improves click performance by four times, which is not an insignificant amount of added consumer attention on your brand.

Feature image courtesy of Shutterstock user Jojje.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

How much better is GPS over Wi-Fi positioning? Yelp knows

Kevin C. Tofel — Fri, 17 Aug 2012 19:38:02 +0000

When you have millions of location data points, figuring out how much more accurate GPS is compared to Wi-Fi is a snap. Luckily, Yelp just happens to have that type of data, captured through users of its app who check-in at various business locations.

Earlier this week, the company scoured the data and shared the results in a blog post. It turns out GPS is twice as accurate than Wi-Fi for very fine data and, possibly due to low-end devices with inexpensive GPS chips, Android devices as a whole are less accurate than iPhones.

Yelp took a deep dive into the geolocation data from June and ended up with a nice chart showing the difference in accuracy from check-ins on the iPhone, Android devices and the iPod touch, which has no GPS and therefore relies on Wi-Fi-based location data.

As expected, the GPS devices are far more accurate than a Wi-Fi only device. Nearly 60 percent of all Yelp check-ins were within 0.1 miles of the actual location on iPhones in June, while just 30 percent of iPod touch check-ins were as accurate. Surprisingly, the Android devices were closer in accuracy to an iPod touch, even though nearly all Android devices have GPS chips.

If you can live with some tolerance for location data, Wi-Fi isn’t terrible: Yelp’s data shows that 79 percent of all iPod touch check-ins were within a mile of the business location.

That doesn’t sound like it’s good enough to ever displace GPS when out and about, but Wi-Fi certainly works better for indoor location needs such as in a mall, large retail store or educational facilities. As a result, both types of geolocation methods are likely to be used in a hybrid form for some time to come.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Skyhook offers devs battery-friendly persistent location

Ryan Kim — Wed, 18 Jul 2012 12:00:18 +0000

For all the talk of persistant location apps like Highlight, Glancee, Banjo and others changing the way we live, many consumers haven’t jumped on board, in part because the battery hit to using these apps has been too great. Location data provider Skyhook said it now has an update to its mobile SDK for Android that will provide persistent background location tracking with no or negligible impact on battery life.

Skyhook’s new Always-On feature, one part of its latest 4.6 version, allows developers to grab location data as frequently as every 30 seconds with accuracy down to 10-20 meters. Ted Morgan, Skyhook’s CEO declined to explain exactly how the company is pulling this off. But he said Skyhook is building off of previous location tracking work it’s done, which downloads more data up front and processes more of the location tracking on the phone.

The previous service, which tried to optimize bandwidth, CPU processing and radios, still meant losing an hour or two of battery life over the course of a day. Now, developers don’t have to deal with any trade-offs, he said.

“We’re now getting more intimate with the way the phone is working and we’ve tuned the system to take advantage of all that,” Morgan said.

This could be a big help for apps that want to surface local offers, highlight nearby services or aid in local people discovery. Now, the apps that use Skyhook have a better shot at delivering real location services without a major sacrifice to the overall user experience. Skyhook is currently deployed on a couple hundred apps that reach 100 million devices.

Developers of location-based apps are getting more help from the likes of Skyhook and, also, Geoloqi, a Portland, Ore. startup that also offers background location processing with very light battery usage. Geoloqi’s SDK works on iOS, something Skyhook isn’t able to do. But Morgan hopes that Apple will allow Skyhook as it sees the need to enable more background location processing.

Skyhook’s updated SDK also offers location tracking for users in airplanes. So apps using Skyhook can determine a person’s location as they fly across the country. That might give rise to some virtual pilot apps that let people know what they’re flying over and nearby points of interest. It could also be a fun addition for location-sharing apps.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

How Facebook made it possible to geo-tag everything

Derrick Harris — Fri, 09 Mar 2012 18:11:53 +0000

It all seems so easy: You log into Facebook, update your status, tell everyone where you are and — voila! — your Timeline is geospatial. You’re not only showing your friends what you did or what you were thinking, your timeline map also shows them where you were when you snapped that photo or felt that particular way. Only, while it’s just one extra step for you to add location, building that capability was a tad more complicated for Facebook.

In a blog post on the Facebook Engineering page today, Facebook’s Karan Mangla describes how it took some serious big-data and architectural handiwork to add broad geo-tagging capabilities. It was actually relatively easy to find locations for Facebook Places when geo-location was only possible with updates done via mobile devices broadcasting GPS signals, but letting users add such data to any updates — even to do so retroactively — meant opening up the range of possible places beyond a user’s current physical location.

Dealing with a larger data set for every update meant Facebook had to change its infrastructure in order to make each update as efficient as possible, lest it risk a slow user experience. Here’s how Mangla said it accomplished this at the database layer:

Depending on the use case, results can be biased toward places closer to the user or provide completely global search. Since the candidate set of places increases significantly in this new system, ranking each of these places to find those that are most relevant to the current user becomes very important. With each place in the index, we store a large set of features to improve ranking, including check-ins and likes received by the place and our best estimate for the opening/closing time for the place. Then we run models generated by machine learning to select the most relevant places based on these features.

With any complex platform such as Facebook, though, there’s always a catch. Mangla noted that while displaying chronological information only requires a server to grab the last few dozen pieces of a user’s activity, displaying geospatial information requires grabbing all the activity associated with any given place on the user’s map. Facebook had a fix:

To manage this data load, we created infrastructure to farm out data fetching to multiple servers. On every page load, a single server fetches the IDs of all pieces of content that can be displayed for the current user. This server then breaks up this data into smaller chunks, and each chunk is sent in a request to another server to actually fetch the data and do privacy checks. The responses from these servers is then combined to create the timeline map display.

The rest of Mangla’s post provides some deeper insights into how Facebook went about predicting locations for previously uploaded photos, as well as its plan for launching APIs that will let third-party apps update users’ timeline maps or feed users relevant information based on where they are at any given time. It’s well worth reading.

The bigger picture, though, is the further confirmation of how much work goes into making a web platform as sticky as possible to ensure maximum user and developer engagement. That means not only adding new features, but also making them perform as well as possible. That’s why, as I’ve written before, anyone expecting to buy Facebook stock and get rich quick might be in for a surprise. It takes a lot of time and money to get this stuff right, and Facebook plans to spend whatever it takes to keep itself relevant over the long haul.

Of course, assuming that all of this location data will eventually make its way to advertisers who’ll then be better able to target users with relevant ads, Facebook should still bring in enough money to keep everyone happy.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Skin Scan wants to fight cancer using iPhones and big data

Derrick Harris — Fri, 03 Feb 2012 18:06:07 +0000

Skin Scan, a startup based in Bucharest, Romania, is selling an iPhone app that lets consumers take pictures of questionable moles and get back information on the likelihood that the mole in question is dangerous. However, during a discussion with Skin Scan’s founders at IBM’s SmartCamp competition on Thursday (where Skin Scan was among nine finalists), I found out the company has much grander goals than helping consumers figure out whether to see a doctor.

On its surface, Skin Scan’s app operatesas one might expect: users take a picture of a mole; the photo is sent to Skin Scan’s servers; Skin Scan’s algorithm analyzes the image; and results are sent back to the user. The app won’t diagnose any condition, but will visually point out abnormalities and will rate moles from low-risk to high-risk. It also refers users to nearby doctors. It’s Skin Scan’s mobile delivery model that makes it so potentially valuable, though.

As co-founder and CEO Victor Anastasiu explained to me, the company is building an analytic database to help make sense of the information it’s uncovering. Using users’ location data, for example, Skin Scan can map the world based on frequency or severity of lesions. Over time, Anastasiu said, Skin Scan should be able to determine how rates are improving or worsening, which is important because skin cancer is often best analyzed over time.

Building time-space models based on mobile data is nothing new, of course. Companies such as Google and Apple are already using anonymous location data to map traffic flows, and another SmartCamp finalist, BitCarrier, is peddling a system to city governments that lets them see traffic flow in real time based on wireless location data and react accordingly.

But Skin Scan has grander plans than even a database. Mircea Popa, another Skin Scan employee, brought up the possibility of disrupting the dermatology system as a whole. One of the company’s next steps is to digitally connect doctors and users via its platform. If doctors can examine patients’ moles without in-person visits, it saves everybody time and money. For offices that are particularly overbooked, Popa thinks Skin Scan could get them to the point where they see only the most-serious cases via office visits.

These are long-term goals, though. For now, Skin Scan is working to gain enough users to get its algorithms as accurate as possible (they’re about 70 percent accurate in assessing severity, Anastasiu said, compared with about 85 percent by dermatologists) and to build a meaningful global data set. Whether or not Skin Scan succeeds, though, its ambitions should resonate with others who want to effect change by leveraging the global reach and broad accessibility of mobile devices.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Foursquare launches version 4.0 with real-time geolocation alerts

Colleen Taylor — Wed, 12 Oct 2011 17:33:57 +0000

Screenshot of Foursquare 4.0 Radar (click to enlarge)

Foursquare on Wednesday launched a new version of its mobile application for the iPhone, Foursquare 4.0.

The biggest change in Foursquare 4.0 is the addition of an opt-in feature called Radar, which is made possible by the latest version of Apple’s iOS mobile operating system, iOS 5.

Radar sends notifications to users based on where they are to remind them to check-in. These alerts are triggered in several ways: when users are close to a place they’ve added to their To-Do list, when they’re close to a place that’s on a Foursquare list they follow, or when three or more friends are checked into a nearby venue. The Radar feature works whether the Foursquare app is open or not, the company said in a blog post.

Radar is an interesting extension of Foursquare’s lists feature, which made its debut in August. Now that the app knows about places a user hopes to visit, vs only knowing about places she’s visited in the past, a feature like Radar can be useful instead of annoying. In all, it seems like a smart evolution for Foursquare, which has been on a growth and R&D tear since it secured $50 million in new venture capital in June.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

Is Groupon a bubble stock or the new Amazon?

Mathew Ingram — Wed, 07 Sep 2011 04:25:01 +0000

Ever since it emerged from Chicago’s small startup community in 2008, Groupon has had nothing short of a spectacular story in terms of its growth: With estimated annual revenues of more than $4 billion after just three years of existence, the poster child for the “group buying” phenomenon is now one of the fastest-growing companies in recent memory. But with the company now reconsidering its IPO and its security filings raising questions, whether or not Groupon’s business model will ever become profitable remains an unknown.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.