Follow Datanami:
September 3, 2013

The Big Potential of that Little Data Miner in Your Pocket

Alex Woodie

In just a few short years, mobile phones have become an irreplaceable part of our daily lives. But the next few years could see phones having an even bigger impact on our lives, thanks to improvements in the area of context-aware computing and the capabilities of phones to extract potentially useful data about their owners, including location, movement, intentions, and whether you’re having a heart attack.

Companies have traditionally shaped their marketing pitches by targeting specific demographic groups. A campaign for arthritis relief cream, for example, probably won’t resonate with Generation Z — the so called “Internet Generation” of young people today. Instead of running ads in Hipster Monthly, the evening TV news broadcasts would work better, thank you.

However, that approach to marketing is changing thanks to the tremendous capability to collect information about potential customers through the various sensors and receivers embedded into smart phones including GPS, Wi-Fi, gyroscopes, and accelerometers. Instead of lumping people into the major demographic buckets — age, sex, ethnicity — companies will soon have the ability to target you based on where you are, what you’re doing, and even how you’re feeling.

The approach is called “context-aware computing” and it’s going to be big. While the term has been around for years, the technology is reaching a tipping point that could unleash all types of potential uses. In a 2011 report, Gartner called it a “game changing opportunity for enterprises to improve both productivity and profits.” 

Targeted marketing is the most obvious (and potentially most lucrative) use for context-aware computing with mobile devices. For example, say a person is waiting to get a seat at a restaurant, which can be deduced by reading GPS and/or WiFi signals, and overlaying that information on an up-to-date, digitized map of businesses in an area. A competing restaurant could use that information to send a customized offer to that person.

However, according to a recent paper on the topic by HTC’s Edward Chang, there are other promising potential uses of context-aware computing. These include improvements in the areas of: information retrieval, facility management, productivity enhancement, power management, and health care.

A phone can conserve power by turning off its GPS receiver when the phone is taken indoors, where GPS typically performs poorly, Chang writes. Similarly, a phone that can detect EKG signals could automatically send an alert to a doctor if it detects an abnormal signal from its user, possibly indicating a heart attack or other life-threatening condition. (Companies like AliveCor make portable EKG monitors that plug onto the back of iPhones.)

The potential uses of context-aware computing are very broad. However, there are technical challenges that must be overcome before it comes into widespread use, Chang writes in his paper titled, “Context-Aware Computing: Opportunities and Open Issues.” In his paper, Chang describes two subroutines that could be used with context-aware computing, including “indoor localization” and “transportation mode detection.”

Indoor localization can even be used to detect a person’s intention, with some degree of accuracy. According to Chang, if a person is walking through a shopping mall slowly (as detected and relayed via the smart phone the person is carrying) that could indicate the person is closely inspecting merchandise, and therefore may be receptive to receiving a coupon. On the other hand, if the person is quickly walking through the mall, that could indicate the person isn’t intending to buy anything soon, Chang writes.

There are other, practical challenges to overcome before context-aware computing can be widely implemented, including problems with “signal variance.” Chang writes, “Sensors manufactured by different vendors may exhibit slightly different error characteristics. Users of different genders, religions, cultures, and builds may move differently. Road conditions and vehicle types/brands can also contribute to signal variance. Cross calibration is required to reduce such variances.”

Phone manufacturers (like HTC) and phone companies are devising all kinds of clever techniques to overcome signal variances, and to correctly determine what people are up to. Another challenge they’ve overcome includes secondary signals that interfere with the primary signal. For example, if a person is holding the phone steady while walking (because she is using its mapping function), the phone’s shock sensor (accelerometer) and gyroscope may not detect the swinging motions of the person’s hands and feet.

Chang writes, “Gyroscopes and accelerometers have been successfully used in submarines and missiles to facilitate dead reckoning. However, the movement of mobile phones and wearable devices is typically not smooth compared with that of submarines and missiles.” Thus gave rise to the Vibration Energy Model (VEM), which uses the signal of swinging energy as the key hint to predict the moving direction. 

Chang also recommends using crowd-sourcing to solve some of the indoor localization challenges. “Given an indoor map, crowd sourcing plots pedestrians’ walking paths, and at the same time, collects WiFi fingerprints at walk-by locations on the map,” he writes. “With sufficient number of collected walking paths, a heat map of fingerprints can be created by performing averaging and outlier removals.” 

Crowdsourcing can also be used to develop points of interest (POI) on a map, and even to predict what a user will do at a POI. For example, given a POI, context-aware computing can predict if a user is shopping, eating, or perhaps waiting for friends, as indicated by pacing. 

Our smartphones have become critical tools in our daily lives. Thanks to the developments being made in context-awareness, they’re becoming even more involved with our interactions with the world.

Related Articles

Data Volumes Strain BI Resources of Online Gaming Firms

The Chief Data Officer’s Time Has Come

Data Science Back to School

Datanami