20 August 2012
There is a dramatic revolution happening in the way maps are made. Geospatial technologist at Google, Ed Parsons, says paper is on the way out as we turn to our digital devices for geographical information.
“It drives me bonkers,” says Ed Parsons, “that every August there are all these articles published moaning about how people don’t know how to read maps anymore.” But according to Parsons, who attributes this annual phenomenon to nothing more than the “silly season”, today we have more access to mapping data than ever before. He says: “We are living in the Golden Age of Cartography and it is geospatial technology that is leading the way.”
Parsons is uniquely qualified to make this statement. His business card tells me I’m in the presence of a ‘geospatial technologist’ at the global information company Google. He is based in the company’s central London office and before we get down to the business of why we are at such a momentous point in the history of map-making, there is the important matter of the famous Google lunch. It’s 4 July – Independence Day – and the Google restaurant is offering hot dogs and turkey roasts. It’s a cultural thing: Google looks after its employees. Beyond giving its engineers the best technology on which to work, the US giant feeds them, gives them ping-pong tables and even washes their clothes. The Google acculturation process is evident everywhere. The idea is simple: if you get the best people and make them happy, then these people will deliver their best. Parsons tells me about the talent recruitment philosophy that is in part a peer approval selection system. “Most traditional HR people would probably think this is crazy. But it works.”
There’s an urban myth that Google’s mission statement is “don’t be evil”. This is shored up by its inclusion in the second paragraph of the company’s Wikipedia entry. (Interestingly, if you Google the word ‘Google’, the Wiki entry comes up second in the Google rankings after a link taking you back to the search page.) The official statement is, of course, “to organize the world’s information and make it universally accessible and useful”.
Geospatial technology is a new term for something that we’re already familiar with: the computerisation of mapping. “That’s something which for the past two or three decades has been known as geographical information systems (GIS).”
But geospatial technology, according to Parsons, has a broader definition and is about “making use of the geographic information that is often created by GIS, but has far more mainstream applications. Yes, it’s still used by enterprise agencies such as utilities companies, or those planning the location of supermarkets and hospitals. But it is also used by individuals when we come out of a tube station and want to know where Starbucks is.” Geospatial technology is about making this data more accessible and Parsons’ role as a geospatial technologist is to be the window into that world.
This window works both ways, says Parsons, who describes his job as painting the vision for Google Maps, Google Earth and “other applications that might run on your mobile phone”.
“The other way to look at this,” he says, “is that we advise agencies such as Travel for London (TfL) about their data and how it could be integrated into Google Maps or assist them to put it into an application programming interface (API) that makes use of geography.”
Parsons describes a scenario where he leaves the building, pulls out his mobile and immediately an application tells him what time the next bus is, or how long it will take him to get home by tube. “The big insight is that the information we are talking about is data that is not on the original paper map.”
On traditional maps, such as the Ordnance Survey you may take with you while out rambling, the data will be static. “You’ll come to a village and your map will tell you that there are six pubs. But it won’t tell you the name of the pub or its phone number. It won’t tell you what local ales it serves, or which of your friends have visited. That’s the sort of information that Google wants to make available. This is the sort of information that you could never put on a paper map.”
With its additional data layers and the capacity for the user to interrogate points on the map, geospatial technology is a broader application than GIS. “If you think of GIS in the financial industry being used for building models, then geospatial technology is more like a spreadsheet – something we all can use.”
According to Parsons, if you were to talk to developers within the GIS community you’d soon discover that there has always been awareness that there is real value in the technology. “But we’ve just not been able to convince people that this is the case.” This perceived image problem is currently being overturned as digital literacy finds its way to the everyday consumer. Parsons agrees: “GIS has had an image problem – a language problem, even – where we developed our own monoculture, self-referentially describing what we do as special and important. This has tended to disenfranchise people outside the community. But now everyone has access to this type of mapping because it is almost completely ambient. Maps are everywhere.”
A journey in the world of maps
Parson’s career kicked off as a geography undergraduate at Kingston University London (or “Kingston Poly as it was then”) in the 1980s. He thinks that he might be the product of a ‘unique time’ because his career spans digital mapping in its entirety. He remembers with wry affection the appearance of the first microcomputers, such as the BBC Micro, Sinclair Spectrum and Cambridge Computer Z88 in schools. According to Parson, the teachers had “no idea what these things were, and so the geeky kids (including myself) would spend their lunchtimes hacking computers”.
The interesting thing about being in his mid-40s is that Parsons has “seen everything from the ground up. Most of my fifth form was spent building a Star Trek multi-player game in Basic. We had a very rudimentary form of serial networking, which meant you could have two players playing the same game. This was a geospatial application. It was like a big version of Battleships where you defined your space in three dimensions and flew off in your spaceship”.
Although he was never taught computer science formally, Parsons “kind of knew about computing”. He became an academic geographer at a time when the first automated mapping was entering the undergraduate curriculum using applications that had been developed at the University of Edinburgh and Harvard. “It was very hand-cranking,” recalls Parsons, who says that if you wanted to build a map, you had to record co-ordinates on paper and then laboriously type them in. “It took forever. But you got a real grounding.”
After graduating, an advertisement in the Evening Standard led Parsons to a marketing company working on postcode boundary maps, where he wrote his own digitising code in Fortran. A year later, he took a masters degree in environmental remote sensing at Cranfield Institute of Technology, following which he returned to Kingston to teach the first undergraduate course in GIS in Europe. This was in the late 1990s, but the importance of this was that it gave Parsons a network of people who are now in the professional geospatial technology career space.
Moving on, and after a short stint in the States, Parsons became CTO at Ordnance Survey (OS) where he spent six years “trying to move a company that was still conceptually about making paper maps to an organisation that was building databases from which you could then make maps. This sounds straightforward but it is actually very complicated both from a technology point of view and a mindset where collecting data for a paper map is different from collecting it for a database”.
In late 2006, Parsons fell out with the management at OS and left in a split that was sufficiently high profile to be reported in the Guardian. There was a carefully worded joint press statement that is easily decoded, after which Parsons wrote on his blog: “I am not in a position to add any more to this statement. But of course I am sorry to be leaving a great group of very committed GI professionals. The future for me is not completely clear at this point.”
This lack of clarity was short-lived as the following day Google telephoned Parsons to ask if he’d like to go to work for them. “And that is not something you easily turn down.”
Mapping the future
“I never really believe anyone who tells me that they’ve planned their career,” says Parsons. “But in geospatial technology there is an underlying issue” that made it hard to hit the trajectory that Parsons has managed to achieve. A career in geospatial technology requires a combination of technical skills that are not taught as an integrated package in any meaningful way. Parsons recognises that there are excellent postgraduate courses in GIS. “But these tend to be very focused on geography, which isn’t a bad thing. But you don’t tend to get the computer skills that you need.” On the other side of the coin, there are intensive computer science courses that fail to provide the geographical knowledge needed. “It’s a hard mix to get right,” he says.
But for those getting it right there are exciting times ahead. Parsons predicts that the main theme to emerge in his field will be increased personalisation of information. Phase one is over: that was the age of taking paper maps and creating automated processes where these could be reproduced digitally. “But this is still related to the notion that the final medium for a map is paper. And paper has restrictions. Once you’ve printed, that’s it. It’s fixed. But that’s not the case with pixels on a screen.”
Parsons believes that we need to rethink cartography in the digital environment. He thinks that a good start has been made, but there’s a long way to go. “Potentially in the future no two maps will be the same. They will be customised for the specific task that the user wants to carry out. This has presented a huge shift in what technology is required to deliver. Today we are producing a billion maps a week. But what we want to do is make each one of these unique.”
The technology challenges behind this, Parsons admits, are incredible. But he is optimistic that we can do it. The market may not have demanded digital mapping to be available 24/7 on mobile devices, but now that it is starting to become available, the market appetite seems to be insatiable.
Crucially, the next generation of maps on our digital devices might not even be maps as we know them, rather a set of geospatial data that has been tailored for the end user. There are now apps with Google Earth integration that can present data – weather conditions, calendar events, and so on – contextually relevant to the user, without being asked.
That’s the end of maps, then? “Maps won’t disappear. But the context will become one of the overriding matters of importance. So, if I’m driving from Cambridge to London on the M11, the big question is always which way to go when I hit the M25: I just need to know – based on current information related to traffic conditions, roadworks and weather – what’s the best choice: left or right.”
Geospatial technology: Tobler’s first law of geography
“We look at geospatial technology as one way of organising information,” says Ed Parsons, referring to the famous Google mission statement. He’s keen to point out that this viewpoint has its roots in Waldo Tobler’s first law of geography that states: “Everything is related to everything else, but near things are more related than distant things.”
What this means in terms of geospatial technology is that when a consumer in Victoria types the word ‘plumber’ into their search engine, then the list of plumbers generated should be local to the enquirer, because “that’s far more interesting. And it’s more useful than plumbers in Glasgow, Swansea or New York”. But to provide this specific list there’s a huge amount of “invisible technology behind the scenes. In this case study, the first thing I need to know is where you are. So unless you’ve told me explicitly that you are in Victoria, I need to be able to find out where your computer is and we can get that from your IP address. If it’s a mobile device then we’ll usually have quite a good idea where it is.
“Once I know where you are, I then need to know where all the plumbers are and then join all that information in about 30-40 milliseconds. And this happens every time someone makes a Google query for anything. We look at the terms and ask whether it’s geographical, and further to that, whether it is a search with which geography can help.”