The Robinson projection is a map projection designed in 1963 by the American cartographer Arthur H. Robinson to display the entire world on a flat sheet in a way that “looks right” without brutally distorting any single property of the globe. Unlike most map projections, which start from a mathematical formula and accept whatever visual result comes out, Robinson worked backward: he decided what the map should look like first, then built a numerical table to produce that appearance. The result is a compromise projection that became one of the most widely used formats for world maps in atlases, classrooms, and international scientific reports.
Why Every Flat Map Distorts Something
A globe is curved. A page is flat. Flattening a sphere onto a plane always sacrifices something, and the cartographer’s job is choosing what to sacrifice. Some projections preserve the true shapes of landmasses at the cost of wildly distorting their relative sizes. Others preserve area faithfully but warp shapes until familiar continents become unrecognizable. Still others keep distances accurate along certain lines but let everything else slide. No projection can preserve shape, area, distance, and direction all at once; the mathematics simply will not allow it.
Robinson’s goal was not to perfectly preserve any one of those properties. Instead, he wanted a map where all four types of distortion were kept moderate everywhere, so that nothing on the map looked grotesquely wrong. Cartographers call this a “compromise” projection, and it is the defining philosophy behind the Robinson design.
How Robinson Built the Projection
Most map projections are defined by a set of equations. You feed in a latitude and longitude, the formula crunches the numbers, and out comes a position on the flat map. Robinson took a different approach. Working at the University of Wisconsin–Madison under a commission from Rand McNally, he essentially sketched what he thought a good-looking world map should be and then reverse-engineered the coordinates needed to produce it.
The result was not a formula but a lookup table. Robinson specified two values at each five-degree interval of latitude from the equator to the pole. One value, sometimes labeled PLEN, controls the length of a given parallel (the horizontal line for that latitude). The other value, sometimes labeled PDFE, controls the vertical distance of that parallel from the equator. Points falling between the five-degree intervals are found by interpolation, essentially by smoothly filling in the gaps between the tabulated values.
This table-driven approach was unusual for the time and initially drew criticism from mathematicians who preferred elegant closed-form equations. But Robinson argued that designing a projection to look right was a legitimate cartographic objective, and the table gave him fine-grained control over every latitude band without locking him into whatever shape a formula happened to produce. Later researchers did develop polynomial approximations of Robinson’s table so that software could compute coordinates without performing interpolation, but the underlying design philosophy remained the same: appearance first, math second.
What the Map Looks Like
If you have seen a world map in an American atlas from the late 1980s or 1990s, you have probably seen the Robinson projection. Its outline is an oval-like rectangle with gently curved edges on the left and right sides. The top and bottom edges, representing the poles, are straight horizontal lines rather than points. This is one of the projection’s most recognizable features: the North Pole and South Pole each appear as a line about 53 percent the length of the equator, rather than as single dots or as infinitely stretched horizontal bands.
Meridians (the vertical lines running from pole to pole) are evenly spaced along the equator and curve gently inward as they approach the poles. Parallels (the horizontal latitude lines) are straight and evenly spaced near the equator but become slightly more compressed toward the poles. The overall effect is a map that feels balanced and natural. Africa and South America retain recognizable proportions. Greenland, which on a Mercator map balloons to the apparent size of Africa, appears much closer to its true relative size. Antarctica, stretched into an infinite band on a Mercator, shows up as a wide but finite strip along the bottom.
Where the Distortions Live
Because the Robinson projection is a compromise, distortion is spread around rather than eliminated in any one place. Understanding where it concentrates helps you read a Robinson map with the right expectations.
- Near the equator: Distortion is minimal. Shapes, areas, and distances are all close to their true globe values in the tropical band. This is the projection’s sweet spot.
- Mid-latitudes: Some stretching appears, but it remains moderate. Countries like France, Japan, or the United States look reasonably accurate in both shape and size.
- High latitudes: This is where the trade-offs become visible. Because the poles are rendered as lines rather than points, landmasses near the poles get stretched horizontally. Greenland and Antarctica appear somewhat wider and flatter than they really are. The area exaggeration is far less dramatic than on a Mercator map, but it is still present.
- Corners of the map: The worst distortion appears in the outermost corners, where high latitudes meet the edges of the map. Regions near the northeastern or northwestern edges, like far eastern Siberia or northern Canada at the map’s margins, are pulled and warped more than anything else on the map.
The projection is neither conformal (shape-preserving at every point) nor equal-area (size-preserving everywhere). It sacrifices perfection in both categories in exchange for keeping both kinds of error tolerable across most of the map. For thematic maps that need to convey a visual sense of the world’s geography without making precise area comparisons, this trade-off works well. For maps where exact area measurement matters, such as a map showing population density per square kilometer, a true equal-area projection would be a better choice.
The National Geographic Adoption
The Robinson projection gained its biggest public platform in 1988, when the National Geographic Society adopted it as its standard projection for world political maps. Before that, National Geographic had used the Van der Grinten projection, which enclosed the world in a circle and inflated the polar regions even more than a Mercator map does. The switch to Robinson was driven by a desire for a more visually honest depiction of relative landmass sizes, particularly in the context of growing public criticism of the way traditional projections exaggerated the size of northern-hemisphere countries at the expense of equatorial regions.
The adoption put the Robinson projection in front of millions of readers and cemented it as the default “world map” in many people’s mental image. For a decade, if you pulled a National Geographic supplement out of the magazine and pinned it to your wall, you were looking at a Robinson projection.
In 1998, National Geographic switched again, this time to the Winkel Tripel projection, which offers a slightly different set of compromises and keeps area distortion a touch lower at the poles. The Robinson projection did not disappear, though. It remained widely used in textbooks, atlases, and scientific publications, and it is still a common choice for anyone who needs a clean, familiar-looking world map.
Modern Use in Science and Policy
The Robinson projection remains a workhorse in scientific communication. The Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES), which produces major international assessments on biodiversity and ecosystem health, adopted the Robinson projection as its standard for global-scale maps. IPBES chose it because it balances distortions in area, direction, and distance while handling the polar regions more gracefully than cylindrical projections do.1Zenodo. Part 1 – Conversion to the Robinson Projection Climate science papers, global biodiversity maps, and UN reports frequently use Robinson-projected maps for the same reasons: the projection is visually intuitive, widely recognized, and does not grossly mislead readers about relative sizes.
In geographic information systems (GIS) software, the Robinson projection is available as a standard option. Software like ArcGIS, QGIS, and the PROJ library all support it natively. The typical implementation defines the projection with a coordinate reference string that specifies the central meridian (usually at zero degrees longitude), uses the WGS84 datum, and outputs coordinates in meters.1Zenodo. Part 1 – Conversion to the Robinson Projection For anyone producing a global map in modern software, selecting the Robinson projection is usually a one-line setting.
Robinson Versus Other Common Projections
The Robinson projection sits in a crowded field, and understanding how it compares to a few of its closest competitors helps clarify what it does well and where others might serve better.
The Mercator projection, probably the most famous of all, preserves angles and shapes locally, which makes it superb for navigation. A straight line on a Mercator map corresponds to a constant compass bearing, which is why sailors loved it. But Mercator inflates areas dramatically as you move away from the equator. Greenland appears roughly the same size as Africa, when in reality Africa is about fourteen times larger. The Robinson projection sacrifices Mercator’s perfect shape preservation to avoid that kind of area exaggeration.
The Gall-Peters projection (also called the Peters projection) goes in the opposite direction. It is an equal-area projection, meaning every region on the map has the correct relative size. Africa and South America finally look as big as they really are compared to Europe and North America. The cost is severe shape distortion: landmasses near the equator get stretched vertically into tall, narrow forms, and the overall map looks unfamiliar and slightly unsettling to people accustomed to Mercator or Robinson-style maps. Robinson trades away Peters’ perfect area accuracy in exchange for shapes that are closer to what people expect.
The Winkel Tripel projection, which replaced Robinson at National Geographic, is also a compromise projection. It averages the coordinates from two other projections to produce a result that, by certain mathematical measures of total distortion, comes out slightly ahead of Robinson. The visual differences between a Winkel Tripel and a Robinson map are subtle to most viewers: the Winkel Tripel’s outline is a bit more rounded, and its parallels curve slightly rather than staying straight. For most practical purposes, the two are nearly interchangeable, and the choice between them often comes down to institutional tradition or aesthetic preference.
The Politics of Projection Choice
Map projections carry political weight whether cartographers intend it or not. The debate intensified in the 1970s and 1980s when Arno Peters, a German historian, promoted his equal-area projection as a corrective to what he called the Eurocentric bias of the Mercator map. Peters argued that Mercator’s size distortion made northern countries look bigger and more important, reinforcing colonial power dynamics. Several international organizations, including some United Nations agencies and aid organizations, adopted the Peters projection in response.
Professional cartographers largely objected, not because they disagreed that Mercator distorted sizes, but because they felt Peters overstated the novelty of his projection (equal-area projections had existed for centuries) and understated the shape distortion his version introduced. The Robinson projection entered this politically charged environment as a kind of diplomatic middle ground: it reduced the size exaggeration that Peters criticized without producing the unfamiliar, elongated shapes that made the Peters map polarizing. The National Geographic Society’s 1988 adoption was, in part, a response to this controversy, a way of acknowledging the legitimate complaint about Mercator’s area distortion while presenting a map that still looked like the world people were accustomed to seeing.
That political dimension has not gone away. Every choice of projection embeds assumptions about what matters. An equal-area projection says accurate size comparison matters most. A conformal projection says accurate local shape matters most. The Robinson projection says visual familiarity and moderate overall distortion matter most. None of these choices is apolitical, and being aware of that is part of reading any world map critically.
Common Misconceptions About the Robinson Projection
One persistent misconception is that the Robinson projection is equal-area. It is not. It reduces area distortion compared to Mercator, but it does not eliminate it. If you need to compare the sizes of two regions on a Robinson map by measuring them with a ruler, your comparison will be off, especially for regions at different latitudes. The projection was designed to look balanced, not to be metrically exact about area.
Another misunderstanding is that Robinson “solved” the distortion problem. No flat map can do that. The Robinson projection just distributes distortion more evenly than most alternatives, so no single region gets badly abused. Readers who see a Robinson map and assume it is an accurate representation of every geographic property are being misled, not by the projection, but by their own expectation that a flat map can faithfully reproduce a sphere.
A third misconception relates to the projection’s creator. Arthur Robinson did not invent this projection as an academic exercise. He was specifically commissioned by Rand McNally to design something that would work well in atlases, where the audience is general readers, not navigators or surveyors. The projection was purpose-built for visual communication, and evaluating it by the standards of navigational accuracy or geodetic precision misses the point of what it was designed to do.
Choosing When to Use It
If you are making a global thematic map, such as a map of internet usage by country, global temperature anomalies, or species distribution, the Robinson projection is a solid default. It gives your audience a familiar-looking world that does not grossly mislead about sizes or shapes. It is especially good for maps that will be printed or displayed at a size where the audience is meant to get a general visual impression rather than make precise measurements.
If your map needs to support area comparisons, choose an equal-area projection instead. Maps showing deforestation rates, population density, or resource extraction per square kilometer benefit from a projection where a given area on the map always represents the same amount of real-world surface. The Mollweide or the equal-area cylindrical projections are better choices in those cases.
For maps of a single continent or a smaller region, you rarely need the Robinson projection at all. Regional maps are better served by projections optimized for the specific area being shown, such as Lambert conformal conic for mid-latitude countries or transverse Mercator for narrow north-south strips. The Robinson projection exists to solve a whole-world problem, and using it for a regional map is like using a wide-angle lens when a normal lens would give you a sharper picture.
For interactive web maps where users pan and zoom, the Web Mercator variant still dominates because of its computational simplicity and the tile-based mapping infrastructure built around it. The Robinson projection is not well suited to slippy web maps because its curved meridians and non-rectangular extent make tiling awkward. Its strength remains what it was built for: a static, full-world view meant to be taken in at a glance.
The Lookup Table That Defines the Shape
For the technically curious, the heart of the Robinson projection is remarkably simple. Robinson published a table of nineteen rows, one for each five-degree step from zero (the equator) to ninety (the pole). Each row gives two numbers. The first controls how long the parallel at that latitude is, expressed as a fraction of the equator’s length. At the equator, the value is 1.0000 (full length). At the pole, it drops to about 0.5322, which is why the pole line is roughly half the width of the equator. The second number controls how far that parallel sits from the equator vertically, also as a fraction of a reference distance.
To plot a point at any latitude, you look up (or interpolate) the two values for that latitude, then use them along with the point’s longitude to calculate its horizontal and vertical position on the flat map. The longitude determines where along the parallel the point falls; the table values determine the shape and spacing of the parallels themselves. There is no trigonometric magic and no complex formula, just a table and linear scaling. That simplicity is part of why the projection has been so easy to implement in software, and it is also why Robinson could design it by eye: tweaking a few numbers in the table shifted the look of the entire map, giving him direct creative control over the final product.