Getting Local: A Map Disassembly

Though it’s only been a couple of months since the last time I did a map breakdown, I feel like it’s time for another. While my own instincts are to avoid repetition, if there’s one thing I’ve learned from my YouTube-watching habits, it’s that a creator can put out a regular stream of content that has little variation, yet remains enjoyable. At least for audiences like me.

Last time, we looked at a map for a popular magazine audience. This time around, the piece is for a small homeowners association, as requested by one of the residents, who is also my brother. He had originally annotated some satellite imagery back in 2009 and shared printouts with others in the community. As I had learned some cartography in the last 14 years, he asked if I’d be willing to help him re-make his older work.

After a few rounds of back and forth, here’s what I landed on (so far; there may yet be some tweaks):

This piece was an interesting challenge because it’s rare for me to work at such a large scale. I’m used to doing small regional maps for books, whereas this is at a nominal scale of only about 1:3,600. I was concerned, before I began, that the result might look fairly empty and boring without something like the satellite imagery of the original piece. But, there was just enough going on in the area to make it work.

Let’s walk through a few highlights of this one.

Color Scheme

Early on, to limit my usual indecision about color, I decided to confine myself to a narrow palette of blue and yellow. Picking two is just easier than having to face the full rainbow of possibilities. These are two colors that pair pleasingly together, especially because the yellow is only at 60%; I think 100% yellow looks too harsh (though it probably prints a bit less harshly than it appears on screen). All blues in the map are tints of 100C 10M 30Y.

These two colors, plus white, allow me to establish a simple three-level visual hierarchy. The yellow draws the most attention (for playgrounds, trails, and other amenities), followed by white (roads/parking), and then the blue background features.

I probably could have made this all work in a blue-white monochrome, but it would definitely be a bit muddier without having the option of pulling some features to the foreground with yellow.

Hypsometric Tint

Since I was concerned about the map being too flat, I wanted to find something for the background other than a solid color. But I didn’t want a distracting, busy satellite image. Fortunately, there was a 1-meter lidar DEM available for this area. Unfortunately, a shaded relief looked pretty terrible.

I wasn’t surprised to find that we’re just too close in to make this work. At this scale, the relief highlights artificial features (even with all the smoothing I applied to the above example), such as the road surfaces. There’s only a few meters of difference between the low and high points in the community, and there aren’t really any landforms to show in this close-up view.

However, I still needed a non-flat background, so I decided instead to do a hypsometric tint. But, a continuous gradient would have had many of the same flaws as a shaded relief. No sense in attempting to show detailed elevation changes where there are none to speak of; it’s only added noise. Plus, given that most of the rest of the map is pretty vector-driven, a continuous raster gradient feels a little mismatched.

Instead, I broke the DEM into four classes using Photoshop’s Posterize tool (I could also have done this in GIS), and produced this:

It gives the impression of elevation change without adding a problematic level of detail. It’s also much more well-defined, vs. the fuzzy raster that, as you saw above, would have taken a much higher-contrast gradient to look like something — a level of contrast which would, in turn, have interfered with the rest of the map content by taking up more of my available color bandwidth.

I did tidy up the edges a bit, by doing some median smoothing in Photoshop and manually removing a couple small zones. I also applied an outer glow to each elevation level, which gave them all better definition. This in turn allowed me to use a reduced range of blue tones to depict them, thus freeing up my palette for higher-priority content.

Glows

There are, true to my usual pattern, glows everywhere else, too (something I blame my friend and mentor, Tanya Buckingham Andersen, for). Basically every polygon or line has some degree of blue glow.

These glows are here for two reasons. The first is the same as with the background tints, above: they help to more sharply delineate the edges of shapes without requiring a high color contrast (c.f. my treatise “On Edges”). You can see the difference when I turn them off. All the layers blend into each other more.

It’s not illegible, but those poorer contrasts are certainly a little less easy to read.

The glows also serve a second (related) function specific to print maps. Maybe I’ll dive into this more in a future post, but the basic idea is: a light feature is made of fewer (or zero) dots of ink, and so a darker glow around it, made of a denser field of ink dots, helps to more clearly delineate its shape. The glows are ensuring that I can get away with white and light yellow text and dashes.

There’s also a glow for the property boundary:

This has a glow on only one side, which creates a sense of inside-outside. The areas on the right appear to recede a bit due to the shadowing of the glow, foregrounding the area on the left as the zone we are interested in. To keep things simple, I did not show any of the boundary, or the glow, along the roadways (where it is implied; parcels usually end at road boundaries).

Blurring

Besides enhancing edges with glows, I also reduced some with blurs (see, again, the previously-linked post on edges). The road labels are foregrounded by softening the road edges underneath.

Compare with an un-blurred version, where the text is legible, but not nearly as pleasant to look at.

You can also see that the houses, which are semi-transparent, benefit from blurring the background tints. In this way, they can avoid looking like they are built over small cliffs. I could have also instead made them wholly opaque, but I wanted them to keep a constant difference between their color and background color, rather than having some houses contrast strongly and some weakly.

All the Textures

Since I’m not using a lot of colors, textures become an important means of distinguishing features. There are various stripes, dots, and hachures throughout.

Each is fine-tuned to the area of interest. For the parking lot stripes (and the coastal hachure on the waterbodies), I’ve manually rotated the pattern for each area to try to avoid potential unsightly messes like this:

While the marsh has a dot fill, I’ve avoided making it uniform. I’ve pulled this trick before, in my Atlas of Great Lakes Islands. By randomizing the size and position of the dots, I achieve just a bit more of an organic feeling.

OSM

This map is built entirely on OpenStreetMap data, though almost none of those data existed when I began. I knew early on that this map would require a lot of tracing of satellite imagery in order to show the fields, buildings, etc. And if I was going to go to that effort, I figured that I might as well trace everything directly into OSM. My brother helped with this, too, especially to adjust some of the tracings in areas that were hard to see in imagery. As a result, his little patch of OSM is far more detailed than the surrounding area.

Wrap-Up

Those, I think, are the primary features of interest. It’s a deceptively simple-looking map of a small area, but it took quite a while and consists of a lot of layers. As usual, good cartographic work inheres in the small details, few of which will ever be noticeable without close inspection.

This will soon be printed out and shared with various people in the community. It’s laid out at 8.5 × 11 inches, but I’ve kept all the text and other important features away from the edges (notice the Bostwick Lake label in the lower right), on the assumption that some will be lost to margins.
That’s all I have for you this time around. Thanks for reading, and I hope you’ll consider aiding my efforts to share knowledge like this by either sharing this post around, or supporting me on Patreon.

PATREON