GIS Cartography - Visual Hierarchy

Visual Hierarchy

All images (and maps are images) have a visual order or depth from background to foreground. A features position in that overall visual order or hierarchy suggests its relative importance to the map. The more important a feature is, the more noticeable it should be and the higher up in the VH it will reside, the less important a feature is, the less noticeable it should be and the further back in the VH it will reside. The distance between features in the VH suggests their relationship to each another (close features are more alike, features further apart are more dissimilar).

Note that the VH in a GIS map is not the same as its drawing order in the table of contents or its real elevation on the ground.

Your VH should follow the intellectual hierarchy for your map. That is, each layers order of importance to the maps purpose should drive how those features appear in the map in order to establish their visual importance to the map user. For instance, in a map of traffic accident data, the points for accidents should be visually more prominent than the symbols of the roads layer they appear on.

from Making Maps - Krygier and Wood

Visual Hierarchy itself relies on a perceptual effect known as figure-ground. In FG our brains classify images and scenes into foreground elements, or figures, and background elements, the ground. Figures in an image are the features our brain says we need to pay attention to and the ground are the parts of an image that we can ignore. Think a little about how your mind interprets the world when driving on the highway versus taking a casual walk in a park and you’ll have some sense of what figure-ground means in terms of our perception. We see differently based on what we perceive as the most important to us at any given time.

We take it for granted that when we look at an image, we know instinctively what part of the picture is important and what is just background.

What is the figure and the ground in this image? How do we know that?

Notice the background is both darker and less sharp compared to the foreground. Within the foreground a clear subject is created by having the darker starfish on a base of lighter skin tones. The subject is also placed near the image center a natural place for us to rest our eyes when scanning the image.

VH is established through changes in symbology, style, and layout. Anything that creates strong contrast or visual difference helps establish the figure-ground relationship. Objects that visually stand out from the background become the figures while objects that fade back become the ground. Figure-ground on a map will fail if the map is too visually homogenous.

Map defaults in a GIS often create a weak visual hierarchy because no single set of features is given more visual weight than any other.

A good fist step to fixing the problem of a weak visual hierarchy is to dispense with any dynamic basemaps which often have too much detail or color for a simple map. Replace the dynamic base layer with your own base features (continents, countries, rivers, lakes etc.) to help set the geographic context. The Natural Earth data are great for setting up simple base maps tailored to your maps design.

Next identify all of the features that are part of the base and establish a unifying symbology that is subdued or devoid of color to help push it to the back of the visual hierarchy. Notice how much easier it is to spot the data points in the 3rd map than it was in the 1st or 2nd.

Solid lines in base layers such as state or national boundary lines can still conflict with higher order features even when thin or greyscale. A great trick for pushing your basemap all the way to the back of the hierarchy while still maintaining clear boundaries it to make all base boundary lines white with grey fill.

Just as repeated outlines can break up the figure-ground relationship between the base and the figure, they can also make symbol reading with in the figure more difficult if there are too many outlines overlapping in a small space. See what your data looks like with outlines that are a darker version of the fill color instead of all black, or even how the symbols may look with out any outlines at all.

In the final example I've taken away all outlines for the data and used the Layer Blending Mode effects in QGIS to multiply the overlapping features, making it easier to see data overlaps and raw data density (will cover better methods of revealing patters in dense data later).