'The Albers USA projection is a composite projection of four Albers projections designed to display the forty-eight lower United States alongside Alaska and Hawaii. Although intended for choropleths, it scales the area of Alaska by a factor of 0.35x (a lie factor of 3), Hawaii is shown at the same scale as the lower forty-eight.' So the projection would affect the information that the map is conveying, though not in a way you might find important. Wouldn't the ideal approach be to have projections change as you scroll and zoom around a map? Zooming into the corner of a conical map means that directions get all skewed. At high enough zoom levels, you Oregon state's Cartography Group had a nice map a few years ago that did exactly this. Optimizing for global viewing when the full world was actually visible and adapting as you zoomed in until you ended up with Mercator at the closest points: We do exactly this on one of our products where I work. Geographic projection with our tiles up to zoom level 12 or so then switching to Mapbox mercator tiles beyond that. Assuming the styling is consistent, the change is more or less imperceptible. Right. I believe this what the National Weather Service does for their radar maps. They're using a polar projection from the location of each station. Points 100 miles north, south, east or west will all fall in a perfect circle around the center of the map. There's much to be said for that. We used that for automatic driving, using the starting point of the trip as the pole. This works well for trips of moderate length. GPS coordinates in Earth-centered, earth-fixed form (a vector from the center of the earth, which what you really get out of GPS) map easily to this form. 'The Albers USA projection is a composite projection of four Albers projections designed to display the forty-eight lower United States alongside Alaska and Hawaii. Although intended for choropleths, it scales the area of Alaska by a factor of 0.35x (a lie factor of 3), Hawaii is shown at the same scale as the lower forty-eight.' The trouble has always been the map data, not the drawing of the tiles. Anyone with a highschool trigonometry education should be able to figure out the drawing given the data. Seems like the point should be the data, not the shape of the map. How does changing the projection type alter the information provided in this case? Yes, you're correct, the point should be the data. However, the projection (not the shape), matters a great deal for a number of reasons. The classic example is to take an orange peel and flatten it. It's impossible to do without tearing or distorting in one way or another. The Earth is of course a 3d shape, and maps are 2d, thus projections to try and map the world as accurately as possible. However, it gets more complicated when you project the data, going back to the distortion. It becomes a problem of 'Pick 2 out of 3.' That usually means Area, Shape and Direction. You can have an accurate representation of Area and Shape, but not Direction. Or in the case of Mercator, Direction and Shape, but not size. Mercator was used by ocean going navigators so that they could draw a straight line on their navigation charts (as longitude and lattitude lines are represented by rectangles with 90 degree angles). Since then of course, we now have thousands of different projections, used for different purposes so that we can try and create the most accurate mapping of the 'real' world as we can. And of course, the projection you choose, should hopefully help you in accurately communicating that geographic information to others. I hope that all makes sense. I think (someone can correct me if I'm wrong) that the Mercator projection doesn't represent the area of the map in a way that accurately reflects the proportions. So, the relative sizes of Texas and Maine shown in the map aren't accurate, relate to each other. Source.