The Unlikely History of the Origins of Modern Maps

GIS technology has opened up new channels of understanding how the world works. But where did it begin?

Swiss map
A Swiss map at the printing office of the Federal Office of Topography in Wabern, Canton of Bern, Switzerland. GAETAN BALLY/Keystone/Corbis

Efforts to explain what GIS actually is almost invariably wax philosophical. At its most essential, GIS is a system for marrying data sets with geography. But it can better be understood as the product of a specific historic moment whose fruit is just coming to bear – a moment arising from the spontaneous amalgam of diverse technologies reaching their apparent apotheosis. And it began when a young Roger Tomlinson—and others—wanted to geographically assess more information than ever before. While the rise of digital culture has served to erode countless boundaries in traditional disciplines, that corrosion partially began in an airplane in 1962 with the predicament of getting gobs of information into one little map.

In May of that year, Tomlinson boarded an airplane flying from Ottawa to Toronto, Canada. He was on a business trip as a 28-year-old geographer for Spartan Air Services, an aerial survey company. Sitting next to him on the flight was Lee Pratt, a government official who had just been named head of the Canada Land Inventory and charged with compiling a map-based catalog of the nation’s productive resources.

Canada may be a large country, but the flight from Ottawa to Toronto is short – a mere hour. Still, in that time, Pratt and Tomlinson struck up a conversation and began chatting about their work. As Tomlinson listened to Pratt describe his plan to collect and synthesize thousands of maps to document the wealth of the vast Canadian landscape, he felt a rush of serendipity. After all, he’d been thinking about the challenge of representing multitudinous data in a map for most of his short career and was on the cusp of programming a computer system for geographic information.

“Do you know how much this is all going to cost?” Tomlinson recalled asking Pratt.

No, was the response – the project was just getting started.

“Okay,” Tomlinson said. “Give me a call if you find that your plan is too expensive to work.”

Three months later, Pratt rang up his fellow plane passenger.


Tomlinson – who died in February at the age of 80 – is now widely credited with conceiving the first functional Geographic Information System, the computer program responsible for the staggering shift in the way we conceptualize spatial information. Decades after Tomlinson’s breakthrough, GIS is everywhere. In late May, the White House announced a deal with the world’s largest GIS firm, Esri, to provide the software free to every K-12 school in the U.S. – an initiative spearheaded by the company’s founder and estimated to be worth more than $1 billion. (Disclosure: partners with GIS firm Esri as an innovative way to tell stories on the site.) The New York Times hires GIS technicians for their digital news desk. Corporations analyze household income demographics and traffic flow with GIS to determine where to expand their retail chains. The city of Covington, Georgia, recently announced an initiative to digitize their local cemeteries using GIS to map online databases of grave markers. 


One of the main problems with maps has always been their dimensionality. In a purely topographical sense, the thing a map is trying to portray is at least one dimension greater than itself. That conundrum has spurred myriad attempts at a solution and cleaved cartographers for hundreds of years – map projections of Earth are still split between ones that say the planet’s a sphere and ones that describe it as an ellipse.

But topography is only one story of space. Imagine a man standing on a road.  Asked to describe where he is in the world, the man might talk about the slope of the hill he’s standing on or the size of the rocks surrounding him. He might also say he’s on a road, between a cornfield and an orchard on one side and a little village on the other. So without even mentioning the birds he observed flying overheard, or the direction of the wind, or the temperature, or even his latitude and longitude, he’s already got six things describing his world. Human experience of space is too complicated to be reduced to the peaks and valleys of a landscape.

This predicament has plagued mapmakers with increasing urgency since at least 1854, when a London doctor named John Snow became dismayed at the relentless onslaught of cholera raging through the city. Snow is now hailed by some as the father of epidemiology for eschewing the commonly held perception that cholera was transmitted by bad air and for painstakingly rooting out the real cause of the epidemic: contaminated drinking water from public pumps. The story looms large in the annals of public health, but just as critical was the way Snow traced the source of the outbreak: by mapping it.

Snow went about the project with medical precision, walking through neighborhoods with a local reverend to document every cholera death by location. They scratched an ink spot for each death on a map, and when they stood back and looked, they saw inks spots clustered around a water pump on Broad Street. Snow’s insistence on the geographic relevance of phenomenon that theretofore had seemed too ephemeral for mapmaking allowed for a new way of seeing.

In an era of sparse data collection, the achievement was revelatory. But data have proliferated ever since, and the question of what to map is limited only by the kinds of number sets people have bothered to gather. The ability to count quantitative variables – computing power – was catalyzed in no small degree by the invention in 1890 of a tabulating machine to numerically code statistics generated by the U.S. Census. Introduced to resolve the processing of large quantities, the tabulating machine, invented by IBM founder Herman Hollerith, eventually became the modern database – the tool that both made a Geographic Information System possible and, by resolving the question of counting, made such a system necessary.


“One of the benefits, if there are any, to being a [former] British colony is that you get a lot of maps,” explained Tomlinson during an interview with Smithsonian before his death. “The Brits are excellent mapmakers.” The abundance of maps holds true for Canada as much as it did for Kenya, and it was part of an international aid effort for the latter former British colony that brought Tomlinson to his fateful meeting with Lee Pratt. Pulp and paper is one of Canada’s most important industries, and the Canadian government decided to share its expertise as part of that aid effort. But Kenya doesn’t have very many trees suited to papermaking. The solution, Canada decided, was to cultivate tree plantations in Kenya, and in 1960, Tomlinson was charged (through a government contract with his company) with finding the best places to put them.

After gathering maps of all the variables he needed to account for in choosing a plantation location – topographic information, demographic information, rainfall patterns, soil quality, atmospheric conditions, animal migration routes – Tomlinson hit an impasse. There were too many.

“When you put six things on your desk and overlay them,” he said, “even when they’re on Mylar sheets, when you start to look down at them you get an awful mess of lines.”

The project was deemed too expensive and abandoned in 1961, but Tomlinson was irked. Computer processing had been making great strides in the late 1950s, and Tomlinson was sure there was a way to put the big new machines to work in assessing all the data. If map areas, called polygons, could be converted to data points and geometrically related to other data points, then each place on a map could contain infinite dimensions of information. The mapmakers’ perennial dilemma of two-dimensional space would dissolve into boundlessness.

Tomlinson started with five polygons on a digital map the size of a cocktail napkin. He put the map on top of another one of equal size, and found that the composite could still make geographic sense. It had solved the isolated variables problem by building a digital sandwich.


When Tomlinson agreed to help Pratt with his land inventory quagmire – a project that resulted in the Canada Geographic Information System, recognized as the first GIS project – they ran headlong into too much data. Looking back, Tomlinson credits his success to the new technological frontiers that transformed computers from overwrought calculators to data storage and processing devices.

Tomlinson and Pratt’s team brought in IBM to help build programming software that could wrangle all the data they were collating. At the same time, Howard Fisher at Harvard was developing a program for synthesizing mapped data at a new laboratory for computer graphics and spatial analysis. In New Haven, Connecticut, the U.S. Census Bureau began to experiment with a system for plotting demographic data by neighborhood block.

By the time Tomlinson finally had his system in hand, which remains in use to inventory Canadian land resources, multiple pockets of innovators had sprung up to make sense of expanding computing power and a desire to understand things in terms of space.

Now in the data age, GIS has since come to influence nearly every sphere of spatial understanding and has helped define a new concept of geography. It’s used to map forest plantations, certainly – but also to track disease outbreaks and to assess demographic changes. Political analysts decipher voting tendencies with it; aid workers predict food shortages with it. There are maps of immigrant remittancesobesityclimate change-induced crisesFoursquare check-ins.

In a 1993 business manual for gleaning financial profit from GIS, an insurance company (that declined to be identified) outlined the company’s policy of refusing to extend insurance in areas prone to certain natural hazards. When Hurricane Andrew hit, the company was not “adversely affected,” because it had simply withheld policies in at-risk locales. 

Some of the most lauded uses of the software, though, include expressly social activist aims. Ushahidi, for example, is a collective and open source mapping effort to document violence that began in the aftermath of the 2007 Kenyan presidential election. The platform, has been used to map violence in South Africa and reports of political corruption in Macedonia. In a nod to history, one of Ushahidi’s most tangible successes was its application in post-earthquake Haiti to determine where help was needed as cholera raged through the island nation.    

The lingering question is whether data-driven maps espouse human creativity or reflect an expansion in the application of computers’ still considerable limitations. Depending on what the wielder of GIS wishes to see, the technology can be enlisted for flattening neighborhoods into formulaic pancakes or extrapolating a collective story. In a map, the primary narrative device becomes not time, but space. As a visual representation, a GIS is like a palimpsest, which, instead of erasing the past to transcribe a new one merely absorbs all moments of the past into an image of the present.

Get the latest History stories in your inbox?

Click to visit our Privacy Statement.