Government agencies turn to Google, Spark data to try and solve Auckland’s traffic woes

The movements of your smartphone are being tracked, collected, marketed and sold to government agencies eager to understand how we get around. National Correspondent Katie Kenny reports.

Mobile network operators have long partnered with big data marketers to sell customer data. In New Zealand, Qrious, a Spark-owned software company, has been providing customers’ location data to local government bodies for the last three years.

Now, it’s experiencing an uptick in demand from central government agencies. Those agencies are also exploring other sources of location data, such as Google and GPS manufacturer TomTom, to help inform decisions and planning.

The Ministry of Business, Innovation and Employment has recently moved from using only official government statistics to incorporating private data.

Its Tourism Data Plan mentions new projects at the Ministry of Transport and the New Zealand Transport Agency, using Qrious and other “travel data” to map drivers’ movements.

The Ministry of Transport says it’s recently completed a project using cellphone tracking data to estimate the volume of travellers between territorial authorities. That data will be used for transport modelling.

NZTA has since confirmed it’s “exploring” using data from Google, which, the search giant says, is sourced using anonymised and aggregated data from Android devices with location services switched on.

WHERE’S THE DATA FROM?

As the types of available data have increased, the costs of obtaining it have decreased. Every time we swipe a bank card, upload a photo to Facebook, use a loyalty card, read an article online, hail a ride with Uber, send an email, we’re giving away some digital information.

Companies such as Facebook, Amazon, and Google mine this information for marketing purposes. But governments, and even charities, can also use big data to try and make our cities smarter and societies fairer.

As of June 2017, New Zealand had 3.8 million mobile phones with active internet connections. Telecommunications companies, such as Spark, see some cellular data for all types of cellphones. (Location data, determined by the nearest cell tower, is necessary for any phone to receive calls.) Telcos can aggregate that data and, as long as they strip it of identifying information, sell it.

While the fine print allows this, most people are likely unaware it happens. For commercial reasons, there’s little transparency in the private sector. Perhaps owing to public trust issues, or the siloed structure of the public sector, there’s little transparency there, either.

When first asked about Auckland Transport’s use of location data, a spokesman simply said the agency didn’t use cellphone data. However, Chris Creighton, group manager of technology solutions, later clarified AT is “considering use” of Google data.

“We’re just exploring these datasets. We have lots of data sources already, so we’re looking at how we can augment this data source to see if it can add any more value.”

The potential benefits include providing drivers with more reliable journey times, as well as measuring the impact of disruptions, he says.

WHAT’S IT BEING USED FOR?

The NZTA says it already uses a variety of data sources to help deliver effective real time travel information to people, and to monitor the performance of the transport system. For example, it purchases historical data from TomTom to “assist with understanding trends in journeys, travel times and speeds.”

Leigh Mitchell, NZTA director for customer experience and behaviour, in an email said the Google data trials relate to providing travel time information to customers through its website and roadside message signs, especially on “key journeys” throughout the Auckland and Wellington regions, as well as improving traffic flow in congested areas.

The data is also being used to help monitor the performance of new state highway infrastructure or upgrades, including measuring travel times through the Waterview Tunnel.

“All Google data is cleaned, anonymised and aggregated before being received or accessed by the NZTA,” she says.

As for how much it costs? “That information is commercially sensitive.

“We use pre-purchased tokens to query Google’s datasets. We do not own the data.”

The results from the Google trials will be compared with existing ways of collecting travel data (such as the use of Bluetooth sensor devices installed around the state highway network, which require ongoing maintenance).

Qrious chief executive Nathalie Morris says the data analytics companyoffers customised reports using “aggregated and anonymised” mobile location data from Spark. This data helps local and central government agencies understand visitor movements, and supports them in planning.

“In particular, tourism and transport industries have long acknowledged there are gaps in traditional data information available, so what we provide is additional insight, allowing an understanding of people flows around New Zealand,” Morris says.

“I think, historically, agencies have tended to use survey data, but what we’re able to provide is a more complete picture.”

The Thames-Coromandel District Council, for example, uses Qrious data to understand population movements over the busy summer period. (Come Christmas time, Coromandel is inundated with Aucklanders.)

“[The council] was keen to look at peak population levels so they could make sure infrastructure is reliable,and understand the volume of people, and where they’re concentrated. So, in the case of an emergency, they know where people are,” Morris says. “That also allows them to do things like support applications for tourism infrastructure funding, event planning, and consent applications.”

Because this data is just coming from Spark customers, Qrious “normalises” it to make it more representative of the full population. As well as location points, the Qrious provides insights such as what part of the country people are coming from (or whether they’re visiting from overseas), and how long they stayed.

PRIVACY VERSUS POTENTIAL

In the commercial sector, there’s an increasing focus on how organisations can use data to drive decision making, and that’s now happening in government agencies too, Morris says.

“There’s a desire to move away from gut feeling decisions to actually be able to use data to drive planning processes. What’s exciting for us is the ability to be part of that, and help public agencies deliver better services.”

There are certainly plenty of advantages when it comes to using this sort of administrative data — that is, data originally collected for another purpose. Administrative data doesn’t involve any extra effort from subjects, such as filling out a survey. And, often with many millions of responses, the insights are more detailed than those of traditional sources. There’s also huge potential for social good.

“There’s huge potential here for getting a much richer understanding of how we’re moving around New Zealand,” says Government Statistician Liz MacPherson. “But the issue is how do we keep the value and maintain trust and confidence at the same time.”

In January 2017, Uber announced it was providing data to cities and transit agencies around the world. The partnership, called Movement, allows city planners and members of the public access to anonymised data from millions of Uber trips in specific cities. Not only can the data alert maintenance workers to a pothole, but it can also measure the impact of major events, such as the London Tower Bridge closure, and analyse movement patterns to help inform urban planning.

As of this month, in the United States, Google is now providing 911 operators with more precise location information from callers using Android phones. (Typically, location is sent to the call taker by a wireless carrier.) The company struck a deal with T-Mobile to send location data straight to 911 call centres. Testing showed the system sent more accurate location data than what carriers provided, and sent it more quickly.

In light of international examples, many technologists would say New Zealand doesn’t go far enough in harnessing the power of big data. However, there are risks to consider, too.

Because it’s originally collected for another purpose, there will always be questions around the accuracy of administrative data used in a different context, and concerns regarding privacy and confidentiality.

In 2013, researchers at the world-renowned Massachusetts Institute of Technology (MIT) studied 15 months of human mobility data for 1.5 million people, and found “even coarse datasets provide little anonymity“.

A simply anonymised data set doesn’t contain names, home addresses, phone numbers or other obvious identifiers. But if an individual’s patterns are unique enough, outside information — public information such as geolocated tweets or pictures or a work address — can be used to link the data back to an individual.

Associate Professor David Eyers, at Otago University’s Department of Computer Science, says sometimes it can be difficult to be sure if data is truly anonymous.

We’re in the early stages of the “data economy”, Eyers says, and it’s anyone’s guess as to how it will evolve.

Given that supermarkets, in some cases, are almost making as much from metadata as from selling products, it would be interesting to know how much telcos get for their data services, he says. But that doesn’t mean the contracts aren’t a good deal for public agencies, he adds.

“Measuring public transport data, that could be hugely useful. To get that quality of data, what would the equivalent instrumentation look like?”

But while consumers remain confused about what data is being used where, they’re hardly in a position to negotiate better deals.

“The data economy is going to end up being about people realising the value of the type of data they have,” Eyers says. It’s “not inconceivable” telcos would drop their prices in exchange for more public data.

“All of this is going to become more transparent over time. It’s just a question of how quickly, and what’s motivating it.”

Stuff