Digital Maps and GIS

Abstract

The computer ate the map twice. First slowly: Roger Tomlinson’s Canada Geographic Information System (1963) invented GIS (geography as a database you can query) and Esri turned it into the quiet trillion-dollar plumbing of utilities, land registries, and epidemiology. Then all at once: Google Maps (February 2005) made the draggable web map a reflex, Street View photographed the world, and free smartphone turn-by-turn navigation (2009) collapsed an entire device industry in a single announcement. In between came the web’s map counterculture: OpenStreetMap (2004), the Wikipedia of geography, now maintained by millions and embedded in half the apps that compete with Google. Maps became the smartphone’s killer app, the location layer became an economy (and a surveillance problem), and cartography (a profession older than writing) became a software feature.

Prehistory: The Map as Analysis

The founding legend of spatial analysis is John Snow’s 1854 cholera map: plotting London’s deaths against water pumps until the Broad Street pump stood accused, the map not as picture but as argument. A century later that idea met the computer. Roger Tomlinson, working in Ottawa on the Canada Land Inventory’s impossible paper problem (overlaying soil, forestry, wildlife, and land-use maps for a continent-sized country), built the Canada Geographic Information System from 1963, the first operational GIS, and the origin of the term. Tomlinson’s insight, which earned him the title “father of GIS,” was overlay analysis: store map layers as data, then compute on their intersections instead of eyeballing acetate sheets.

Academia industrialized it. Howard Fisher’s Harvard Laboratory for Computer Graphics and Spatial Analysis (1965) spread mapping software (SYMAP’s line-printer choropleths, GRID, ODYSSEY) through universities worldwide. A Harvard Lab alumnus, landscape architect Jack Dangermond, founded Esri in Redlands, California in 1969; its ARC/INFO (1982) and successor ArcGIS became the Microsoft Office of the field. The US Census Bureau’s TIGER database (built for the 1990 census) put machine-readable streets for the whole country into the public domain, the substrate on which the consumer mapping industry would later be built. By the 1990s GIS ran land registries, utility networks, redistricting, retail siting, and disease surveillance: enormously consequential, almost invisible.

The Web Gets Directions

MapQuest (February 1996) put driving directions in the browser and became one of the early web’s indispensable utilities (printed MapQuest directions were the era’s navigation medium) and AOL bought it in 1999 for $1.1 billion (see The AOL Era). But web maps were server-rendered: click an arrow, wait for a new page.

The modern map arrived from three 2004 acquisitions Google made in quick succession: Where 2 Technologies (the Rasmussen brothers’ mapping engine, built in Sydney), Keyhole (satellite-imagery flythroughs, partly funded by the CIA’s In-Q-Tel; it became Google Earth, June 2005), and ZipDash (real-time traffic). Google Maps launched February 8, 2005: pre-fetched tiles, AJAX, a map you could drag, alongside Gmail, the demonstration that the browser could feel like an application (see The JavaScript Revolution). The Maps API (June 2005), born partly because developers were reverse-engineering the tiles anyway, made the embeddable map the original web mashup and a pillar of The API Economy. Street View (May 2007) drove camera cars down the world’s roads, a project of Google-scale audacity that also produced Germany’s fiercest privacy backlash and a landmark Wi-Fi-sniffing scandal (see The Privacy War).

The decisive blow was economic. In October 2009 Google announced free turn-by-turn navigation on Android: Garmin and TomTom (companies built on selling exactly that for hundreds of dollars) lost double-digit percentages of their market value within days. Navigation stopped being a product and became a feature; the smartphone map became, with messaging, the most-used application on Earth, over a billion monthly users for Google Maps alone by 2020 (see The Mobile Computing Revolution, and GPS and Satellite Navigation for the satellite layer beneath all of it). Waze, the Israeli crowdsourced-traffic navigator, proved users would be the sensor network; Google bought it in 2013 for ~$1.1 billion rather than face it (see Israel’s Tech Industry).

Even the giants could fail at this. Apple Maps shipped in September 2012 with melting bridges and misplaced cities, produced a public apology from Tim Cook recommending competitors’ products, and contributed to a senior executive’s exit, then, a decade of rebuilding later, became genuinely good: the rare tech disaster with a redemption arc (see Tim Cook and Apple).

OpenStreetMap: The Wikipedia of Geography

In August 2004, London developer Steve Coast, irritated that the taxpayer-funded Ordnance Survey sold its data back to taxpayers under Crown copyright, started OpenStreetMap: anyone with a GPS trace could map their street, and the result would be free forever, like Wikipedia. It looked quixotic until it didn’t. The Haiti earthquake (January 2010) was the proof event: within days, remote volunteers of the Humanitarian OpenStreetMap Team traced satellite imagery into the most detailed map of Port-au-Prince ever made, used on the ground by rescue teams. OSM grew past ten million registered users and became the base layer for much of the non-Google map world: Apple, Meta, Microsoft, and thousands of apps consume it, and the Overture Maps Foundation (2022; Meta, Microsoft, Amazon, TomTom) formed essentially to industrialize open map data rather than keep paying Google. The commons strategy worked the same way it did for Linux: the map everyone needs and no one wants to build alone became shared infrastructure (see The Open Source Revolution).

The Location Layer

The map’s last transformation was becoming invisible: a permission dialog. Ride-hailing, delivery, dating radii, fitness heatmaps, geofenced advertising, find-my-everything, the gig economy is a GIS application, and “location history” became one of the most intimate datasets ever assembled, implicated in everything from the Strava heatmap revealing secret military bases (2018) to law-enforcement geofence warrants querying who was near a crime. Tomlinson’s overlay analysis now runs continuously, on everyone, in real time. The map knows where you are; the policy fight is over who gets to ask it (see Surveillance Capitalism).

⚠️ Dead End: The Standalone Navigator

The personal navigation device (Garmin nüvi, TomTom GO) was one of the fastest-growing consumer electronics categories in history, tens of millions of units a year at its 2007–08 peak, built on a beautiful margin stack: proprietary maps (Navteq, Tele Atlas licensing), $200–500 hardware, paid map updates. The 2009 free-navigation announcement didn’t out-compete the category; it deleted its reason to exist, the phone in every pocket had the screen, the GPS chip, the data connection, and now the software, at a marginal price of zero. Nokia bought Navteq for $8.1 billion (2008) and TomTom bought Tele Atlas for €2.9 billion just in time for the asset to stop being scarce. The instructive detail is who survived: TomTom retreated to licensing data and services; Garmin pivoted to wearables and aviation/marine (places where a dedicated device still beats a phone) and thrives. Same lesson as Wang and the word processor, one generation later: when your product becomes a feature of a general-purpose device, sell the layer that’s still scarce, or die with the box.