I finally spent a bit of time catching up with the great Daniel Huffman’s excellent tutorial on creating shaded relief maps with Blender. Do try it out if you haven’t; he’s written it up thoroughly and helpfully, showing and explaining why Blender is good for this purpose. I’ll wait here while you read through it.

The White Mountains of New Hampshire provide the setting for my attempts. This is where we go from Boston for good hiking and other outdoor activities. My significant other has gotten me into enjoying the area in all seasons. Mostly we do big hikes up various mountains, but she’s even persuaded me to try skiing for the first time in my life soon.

Anyway, an idea came up of making a shaded relief map of the White Mountains in each of the four seasons. Here’s what I’ve got so far. (Click/tap for big version that’s like 27 MB.)

This is a combination of elevation and blurred land cover data, plus TIGER/Line water polygons. Spring, summer, and autumn all use the same hillshade, but winter’s different, having a lower and weaker sun. Besides that, it’s a bunch of Photoshopping to try to evoke the look and feel of each season.

Summer (top right) is probably the most straightforward, with the land cover layer mostly left alone. Green areas are a bit yellow, an attempt at a kind of warm, dry look. Spring (top left) colors are little cooler and brighter, but only at lower elevations. Deciduous forests up higher are still bare, and there’s even still some snow on the higher peaks. Later, in autumn (bottom right), I made a splotchy texture of greens, yellows, oranges, reds, and browns to suggest the lovely fall colors of the region. In winter (bottom left) I really went for it on the lighting. As mentioned earlier, the rendered relief has a low sun. Shadows have a lot more blue than the other seasons, and highlights are warmer, as though the sun were setting. And, of course, there’s snow. In reality the whole region would be snowy, but I liked having some variety, so the lower parts are bare, except evergreen forests. Evergreens, of course are the only thing that’s green across all four maps—not that I was consistent in their color.

Technical tricks are probably not really worth getting into here because I’m not much of an expert. Most of the things I’ve tried involve Photoshop adjustment layers and layer masks based on elevation and/or the shaded relief itself. When the shadow and highlight layers are cranked up to 11 on the winter map, for instance, you can see how the areas to which they apply are based on both lighting and elevation. Daniel Huffman also has some related tips (and many more) in his article on terrain in Photoshop.

The winter map especially showed me how nice Blender is for shaded relief. As you’ll hear from Daniel, it’s 3D modeling software that renders realistic lighting and shadows. When the light source is low, as in my winter map, you can really see how shadows behave like actual shadows. For example, notice how the Presidential Range (left, below) casts a shadow onto the other side of the valley. A more typical hillshade doesn’t really model shadows but rather shades the left side the of valley (facing away from the light source) dark and the right side light.

Anyway, I’ll probably keep messing around with these maps forever, and try to make a real map with labels and everything. Meanwhile, I’ll also keep visiting the White Mountains frequently!

The pages are mostly empty still, but they’ve got a handful of sketches mapping a summer highlight: walking the Camino de Santiago in Spain with my dad and sister.

These maps went into a commemorative photo book that we put together as a Christmas gift to my dad. My sister had done this walk before, in 2012, and sparked an interest in my dad of “some day” doing it, which became “now” before time grew too unkind to one’s ability to walk 800 kilometers. Walking long distances day after day was a lot harder than expected despite my fairly frequent strenuous hiking near home (my feet didn’t stop hurting for a few more months), and my dad especially endured a lot of aches and pains, but he never quit or even took a day off, impressing fellow peregrinos we met along the way. 38 days after departing Saint-Jean-Pied-de-Port in France, we arrived in Santiago de Compostela.

People walk the Camino for a variety of reasons: from making the traditional Catholic pilgrimage, to escaping stress at home, to tourism. We didn’t exactly set out with any grand purpose, but with weeks of walking, often in quiet solitude, one starts earning that “spiritual” checkbox on the form that’s filled out at the end. For me, the trek is most meaningful for the family experience, but the geographer in me had a pretty good time, too.

Not to be a preachy “stop and smell the roses” guy, but it’s true that especially in the day of GPS-enabled navigation, we don’t often get to know the places we move through. Nothing beats endless, slow walking for contemplating your environment. We became well-versed in the cereal crops of northern Spain; we really saw every tree and rock (i.e., shade and rest); we took in the cultural landscape (architecture, people, food etc.) at the pace it deserves.

We used good ol’ paper maps more than I ever do these days, two of them living in the outer side pocket of my backpack, within easy reach. We mostly disabled mobile data for the trip (expensive!), and although I often downloaded a few days’ worth of area on Google Maps when we had evening wifi, it was a lot easier to plan out each day’s walk and breaks using printed maps. Mostly these weren’t for navigating the path, as the Camino Francés is generally very well marked, but for deciding distances and routes (when there was an option).

I drew five maps, divided up by the regions we passed through. They simply show the route of the Camino with our daily stops, along with little sketches of things we saw along the way: landscapes, foods, landmarks, and a few things peculiar to our personal experiences. It’s a pretty gentle way into hand-drawn mapping, mostly being a few lines on top of the notebook’s helpful grid. (The pages also have nice north/south/east/west labels, but I had to ignore them and turn things sideways.) It’s nothing fancy, but it’s progress for my map-drawing and tiny sketch-drawing abilities, which are generally somewhere below kindergarten level.

I could go on and on with like a million photos from the trip, but I’m mostly here to show the carto-mementos. To anyone who may be doing the walk sometime: ¡buen camino!

To recap, these arcs represent straight-line paths out to sea, perpendicular to the coastline at a sample of locations (around 16,000 points). These “views” often lead to surprising destinations, given the flat maps we’re accustomed to seeing.

This interactive map is just a toy that I hope looks nice—there’s nothing especially marvelous on the technical end. The paths were all calculated beforehand the same way as before; while it’s certainly possible to calculate them on the fly, my method isn’t quite fast enough to keep up with rapid interaction. (And I’m too lazy to solve a few other challenges that would come with it.)

The short technical summary is this. Coastline sample points are drawn (invisibly) as SVG circles, and when you mouse over them their associated path is drawn (again, invisibly) as an SVG path via d3. That SVG path is handy because we can get locations incrementally along it with the getPointAtLength method, which is used for drawing the path on a canvas bit by bit in an animation. There is a constantly-running function that draws any active paths, preceded by a barely-opaque fill across the board, which is how the fading trail effect is achieved. Source code is a bit messy at the moment, but is up on GitHub.

One bit of magic is that d3’s geoPath uses adaptive resampling to draw lines as curves appropriate to a map projection. This means that, for the most part, I only had to include the start and end points in my line data, and d3 would draw the curved great circle path that I was looking for. (I did, however, sometimes have to include extra vertices to convince it to go the longer way around the world when necessary.) As you may be able to see in the QGIS screenshot below, it’s all made up of straight lines.

Anyway, poke around at the map and enjoy the light show!

[Edit: an interactive map should appear here, but there are some unresolved issues. Check out the GitHub-hosted version for something at least partly functional!]

What we have here is more or less an animated flowy map of drainage, although it’s not a serious attempt at an accurate map. It’s just something that turned out to be kind of pretty.

Apologies for its clunkiness; I’m not that good at code. This is another episode in my habit of messing around with terrain mapping in the browser on canvas elements. See for example shaded relief and grassy sketchy thingies. (Side note: calculation of aspect turns out to be poor in those; this way works better and I’ll fix them at some point.) None of this ever turns out zippy, but it’s always a good learning experience and is occasionally art-worthy.

Maybe you can guess what’s going on here. I throw a bunch of random points at a terrain map, and they begin flowing toward lower elevation. That’s all, really: each path continuously asks “what adjacent location is lowest?” and moves in that direction. It’s more complicated behind the curtain, of course, but it’s not awful. I’ve tried to leave helpful comments in the code.

I’ve been inspired for a while to attempt this kind of animation by Cameron Beccario’s amazing global map of wind, etc. (perhaps spurred into mind by everybody posting links to the map during Hurricane Matthew) and the original wind map by Martin Wattenberg and Fernanda Viégas, and other similar maps that they have in turn inspired. I’m not as smart as those people and don’t entirely know how they’ve done it, but I was pleased to learn how to do a kind of fading trail animation one way or another.

The key to all this is Mapzen’s terrain tiles. The hillshade you see in the map is straight magic from them, and replaces the things I was trying to do on my own with DEM images and canvas. Invisible but crucial are their Terrarium elevation tiles, which are these wild-looking purple and green things—but which encode actual elevation in their colors. Pretty clever.

Just load those in, do some pixel calculations to find which way is downhill, throw down some paths, and animate those according to the calculated directions. And there, you’ve got yourself a map that’s pretty fun to watch.

Anyway, play with the map, and have a look at the GitHub repository if you want to see the code or tell me about anything dumb I have done.

Silly because, duh, a chart of land area by latitude and longitude is a map. Then again, a map is both latitude and longitude. What if you look at each dimension individually? Heck, let’s just try it.

I did this by, essentially, drawing maps and counting pixels, then collapsing all those pixels against an axis. The important thing is to use an equal-area projection so that each pixel means the same thing everywhere.

For latitude, an equal-area projection with even spacing between parallels so that the axis has a consistent scale of degrees: sinusoidal.

SQUISH!

And similarly, for longitude, one with even spacing between meridians: a cylindrical equal-area projection, Gall-Peters.

MELT!

I’m not sure what there is to gain from these quasi-maps. Maybe nothing. But it’s kind of neat to see how the shapes of some continents are still vaguely there even when all squashed together. (By the way, apologies to places like New Zealand that got lumped together with Australia.) It’s kind of like those layered jars of sand art.

If nothing else, it’s fun to go full Play-Doh and cram everything both ways into a blob down the corner.

Yeah, so satisfying.

I’ve never been a terrain representation expert, but occasionally I get briefly super interested in some DIY technique for relief mapping, not using typical GIS tools or rendering software. Tom Patterson’s old (but still applicable) Photoshop tutorials were my introduction to the magic simplicity of turning a grayscale image into fancy shaded relief.

Some years ago I wrote code for simple shaded relief, using Flash at the time. I picked that up again just a few months ago and did the same using JavaScript and canvas elements. These were impractical experiments, but they were a super interesting way to learn some basics about how automated relief shading works, and they were a more fun playground than turning knobs or filling in boxes in some software would have been.

Then the other day, for no particular reason, hachures came to mind. Hachures are an old-fashioned technique in which parallel lines are drawn along slopes. You don’t see them much these days, but they’re one of my favorite techniques; their chapter is still the only one I’ve managed to read in Imhof’s book. From time to time I’ve wished there were an easy way to generate nice hachure maps in GIS or what have you—as far as I know there’s still not much out there.

I tried rendering some hachure-like maps based on the same code I’d used before, that is, in a browser with just some JavaScript. The ultimate result was not so much hachure maps but the kind of sketchy map style you see in these images, almost like something drawn with colored pencils.

It began with something that I quickly recalled is essentially what Michal Migurski had worked on (later inspiring Eric Fischer) some years back: divide elevation data into a grid and use short strokes to show slope (as line width) and aspect (as rotation).

That was kind of cool, but certainly a lot more rigid-looking than classic hachure maps. The next step was to introduce some small random offsets in the grid so things didn’t line up so perfectly, looking slightly more natural.

Wanting to move even further from a gridded appearance, next I tried drawing longer strokes at each point, longer than the dimensions of the cell so that lines would start to blend together. I also remembered that I wanted shadow hachures, where stroke weight or shade corresponds to brightness in an illuminated scene, not just slope. (My earlier shaded relief code could calculate this.)

At this point it was starting to look more like a pencil sketch than a true hachure map, so I ran with that and added color. There’s kind of a combination of slope and shadow hachuring here, with stroke width representing slope and color representing shadowing.

Lastly, there were two subtle steps: add a little more randomness in the form of slight curves, and draw the map at lower opacity several times on top of itself, each time with the sun angle varying a bit. This produced a somewhat smoother scene with a bit more shadow detail, and allowed the various random effects to accumulate into a very grassy look.

The source code for this is on GitHub here. All in all, it’s not practical code, at least not in this implementation (it would be interesting to try it in a more efficient way, or on something like Mapzen’s elevation tiles), but it’s a fun style to play around with! With more attention to things like color I can imagine some pretty good-looking maps. In any event, here are a few more that I rendered.

Crater Lake:

White Mountains, around Mount Washington:

Oahu:

Yosemite Valley:

Hocking Hills, Ohio:

What’s really across the ocean from you when you look straight out? It’s not always the place you think.

I’m inspired by a map done a couple of years ago by Eric Odenheimer and some follow-ups by Weiyi Cai and Laris Karklis of the Washington Post. Those maps are colorful, handy guides to countries of equivalent latitude across the oceans. It’s easy to forget, for example, that much of Europe is well north of the United States east coast. But they’re not exactly maps of what’s across the ocean from you, at least not directly across from you. To think of east or west as “straight” across is, perhaps, one of those effects of the map projections we see every day.

The latitude maps got me interested in answering the question more strictly: standing on a given point and facing perpendicular to the coast, if you went straight ahead, never turning, where would you end up? There are two reasons why following a line of latitude won’t answer the question.

1. Coastlines are crooked and wacky.
2. The earth is round.

With that in mind, here are some maps showing the points from which you can “see” each of the continents.

Coastline angle

Coastlines face all different directions, bending and turning constantly. The “East coast” isn’t a straight north-south line facing directly east. Just look at the state where I live, which has coastline facing literally all directions.

Taking “across the ocean” to mean directly across, perpendicular to the coast, then what’s across the ocean depends on where you’re standing! To get a rough idea of what direction the world’s coastlines face, I’m calculating the angle between every pair of adjacent coastal vertices in medium scale Natural Earth data, then placing a point in between them and measuring the view from there based on that angle.

The much-maligned Mercator projection comes in handy here. Those angle calculations were made using projected coordinates because the conformal Mercator projection preserves the thing I’m interested in: local angles!

Straight lines on a round object

The second point is trickier to imagine thanks to common rectangular maps and the way latitude itself is defined. If you can detach the concept of “direction” from the concept of east and west, and look at globes and other map projections, it’s easy enough to picture. The shortest, straightest line on a sphere (let’s call the Earth a sphere even though it technically isn’t) is a great circle arc, not something like a line of latitude.

What we often think of as “straight” is a path following a rhumb line, a line of constant bearing. Wikipedia succinctly describes how such a “straight” line actually turns, in contrast to a great circle.

If one were to drive a car along a great circle one would hold the steering wheel fixed, but to follow a rhumb line one would have to turn the wheel, turning it more sharply as the poles are approached.

A typical classroom demonstration of great circles is to pull a piece of string taut on the surface of a globe between two points, and note how the string arcs across lines of latitude, changing its bearing the whole way. Try this specific case to drive home how the spherical “straight” differs from “straight” as we’ve defined compass directions: find a line of longitude on the globe, then a spot along that line somewhere away from the equator. Bring the globe to your eye and place the string perpendicular to the meridian, in between two latitude lines. Line up your view with the string and you can see that even though it starts out going due east or west, as it continues directly ahead the “straight” east/west parallels curve away from it.

So if we want to know what’s truly straight across the ocean from a given coastline point, we need to see what direction the coast faces at that point, then draw a great circle in that direction and see what it runs into.

As for flat maps, certain map projections provide an accurate view of directions. The azimuthal equidistant projection, for example, preserves correct direction (and distance) from the center point of the map. A straight line from the center of the map is a straight line in real life. Here’s the Newfoundland-to-Australia example from earlier:

Such a map, in the end, is how I’m figuring out beach views: center that projection on each point, then draw a straight line in the correct direction until it hits land.

Conclusion

I’m not entirely certain that I have all the math right, but I think it’s at least close. Even we cartographers sometimes have a shaky grasp of map projections and spherical geometry.

But who has time for correct math? I’ve got to start training for the straight-line swim from the number one beach in my life—30th Street in Ocean City, New Jersey—to Brazil.

Anyway, one broken thing was the crazy Ohio piano map I made nearly seven years ago, in which counties of Ohio were represented by the sounds of piano keys, and data became music of sorts. Actually, it’s not entirely broken; the sounds and most interactions work, just with a big, persistent gray error because Google Maps for Flash is no longer a thing.

Professor Robert Roth insists that this ridiculous map is vital to the teaching of cartography, and who am I to stand in the way of education? So I dug up the old laptop that had the only copy of the Flash map’s source code and set about recreating the map in more modern HTML and JavaScript. It doesn’t have everything the old map has, but hey, it works. Bonus: now the code can actually be seen, on GitHub.

It definitely does not work on mobile devices because playing audio elements on mobile is complicated and I don’t care enough, but if you’re on a desktop computer, here you go. Don’t touch this if you’re in a library or something without headphones. You’ll be embarrassed.

In the official Cartographic Summit photo above, I’m the robot on the right edge who doesn’t know how to stand like a normal human. I’ve just returned from this unusual event, which was part traditional conference and part unconference, with people in attendance only by invitation. I didn’t know what to expect, but figured a few days with friends new and old in sunny Redlands, California in February couldn’t hurt.

The point seemed to be to discuss and identify key areas of focus and/or research for cartography in the near future. A handful of excellent speakers got us going, but much of the main event revolved around small discussion groups which met to talk about challenges and ideas along certain themes and keywords, subsequently summarizing the discussions for everyone else to hear. Surprisingly these were very engaging, forward-facing, and dare I say somewhat fruitful. Usually traditional cartographers talking about Cartography Future only want to fix what’s wrong with Cartography Present, and blame it on those who don’t know Cartography Past. Less so this time.

Altogether it was pretty interesting, and I was mostly pleased with what transpired. Below is a short list of words and phrases that stuck in my mind. Some are good and some are bad. A few are my own thoughts but most paraphrase what I heard. Many are open questions, but that’s okay. Part of the idea here was to develop agendas. Do check the #cartosummit hashtag on Twitter for many more reports, and watch for audio and slide recordings to be posted.

AUDIENCE. We were traditionally taught to understand our audience before designing a map, but we may be falling behind these days. We don’t always think enough about how our maps will be used in the wide variety of possible media and environments. And the internet being the internet, sometimes a map escapes our control and goes far beyond its intended audience. Related to this, there was some discussion about making better use of tracking and analytics, and making such capability easier in our tools.

BIG DATA, SMALL MAPPING. Big data isn’t entirely new. People mapped massive data sets ages ago, but in those days it was usually done by large organizations with lots of people and resources. These days the entire job often falls to an individual. What are the consequences of this, and how can we as cartographers address such a challenge?

BLACKLIST. Although this event was hosted by Esri and organized in part by Esri people, it was in large part an ICA event and Esri attendance was very limited. But for an intellectual meeting about the future of cartography, there was a conspicuous absence of representatives from certain major players in mapping—companies that are in some degree of competition with Esri. Rumor was that inviting them was denied by the Esri powers that be. I’ll take it back if that’s not true, but if it is, it sours the experience a bit for me.

CLARIFY. The best-remembered word from the inimitable Nigel Holmes this week. It’s understood that a vital role of a cartographer is to find and make clear some aspect of data, not simply throw all the data at everyone. This was especially emphasized in discussions of mapping “big data.”

COMPUTATION. What is the role of computation in cartography? It’s more than the hard work of crunching through data. We should embrace it as part of the design process, not as something that detracts from it. If there is one thing I would stress to modern big-C Cartographers, it’s this.

DIVERSITY. This was definitely a white male, people-like-me meeting. At the very least, clearly in a group of 50, there ought to be more than 11 women. Diversity is something we need to improve in the cartography community in general, not only in the sense of gender, ethnicity, etc., but also in things like educational and economic background. Most people who actually work with maps don’t come from the world of advanced geo-degrees that we tend to deal with. (Side note: we’re working on some plans for this at NACIS.)

ELITISM. I appreciate an effort to bring together top minds in cartography (plus a few of us internet loudmouths) and the huge benefits of smaller group discussions, but in a field as collegial and small as ours it leaves many people spurned and furthers an elitist reputation we have among some segments. A few poorly contextualized tweets make it look like we’ve only gathered to continue picking on the new kids. I really don’t know how to be limited, representative, and diplomatic all at the same time, but there is probably some ground to gain here.

EXCITEMENT. Cartographers are in an enviable position to make other people excited about their own domains. The smartest people in the world can have their minds blown by seeing something they’ve studied for years mapped for the first time. We can do good for other fields at the same time as our own.

FAST AND GOOD ENOUGH. Let’s recognize that many modern map users and map makers want to accomplish something in mere seconds, not through a slow series of decisions. We can do people a great service by getting as far as possible with smart, automated decisions. The result will never be perfect, but it’s worthwhile.

MAP LITERACY. How can we better teach ordinary map readers to look at maps critically, to understand their limitations and biases rather than taking everything as unqualified truth? There was, at least, some sense that people are getting better at recognizing good cartography, which is nice. I’d rather help people become better map readers than worry about a proliferation of bad maps.

SKIPPING. Yeah, so I ditched most of the third day to go to Joshua Tree National Park with [names redacted to protect professional reputations] instead.

It has become apparent that you are all terrible at being pedestrians. I feel confident saying that despite some good eggs and battle-tested walkers such as New Yorkers, most Americans are obliviously selfish, inconsiderate users of sidewalk space. It’s true even in a city like mine, where a large chunk of people commute on foot.

Perhaps it’s because cars are the default mode of transportation in this country, and once on foot outside a car, people view everything as some kind of personal leisure space, not a system that still requires a certain amount of order, competition, and cooperation.

Whatever the cause, the point is, my friends, that you are always getting in the damn way of other people who are trying to get places. I know you don’t mean to, but you do.

But I’m here to help. Here, using some handy diagrams, is a short guide to not being a jerk and getting the hell out of the way.

You’re welcome. Now get out of the way.