A Holiday Shopping List

When I started my first map-related website back in 2010(?), it had some text on there about how I thought it would be cool to bring together map artists into one place where people could buy our products. This may have been borne out of my own failed attempts to sell a couple maps of my own (things have gotten at least a little better since then).

After this idea sat in my mind for a decade, I finally decided to do something about it. Capitalizing on the holiday shopping season in the United States, I’ve put together a list of small businesses selling map prints and other products. Check it out and spread it around (and support independent map artists)!

(Click that there image to go to the list)

This is also the start, I hope, of something a little larger. I’d like to form a cooperative of independent sellers of map products. I think we are stronger if we band together. If you run a small business selling map products to consumers, feel free to drop on by The Spatial Community (thespatialcommunity.org), which is a public Slack group for, well, people who do spatial stuff. I’ve formed a channel there called #map-products-cooperative. Join the channel and we can start to organize ourselves!

The Quest for the Blue Moon

Where I come from, many ice cream places offer a flavor called blue moon. In the freezer case, its unnatural coloration immediately draws attention amidst the more pedestrian offerings of chocolate, vanilla, and strawberry.

Photo by Bill McChesney, via Wikipedia. CC-BY-2.0

Its actual flavor (and coloration) varies from purveyor to purveyor, each with their own secret recipe. According to Mashed.com, “Some say that the ice cream has a breakfast cereal sweetness similar to Froot Loops, while others take in notes of almond or citrus.” In truth, it’s been so long since I’ve had any, I can’t really remember. All I know is that it’s a standard offering at ice cream shops.

Or so I thought. In recent years, I have come to learn that it’s not widely available throughout the United States. It is, instead, a regional flavor, with various articles describing it as a “Midwestern favorite,” and an “iconic Midwestern frozen treat.” But nothing I read was able to give more detail about where blue moon was found — only anecdotal, unsatisfying generalizations about the Midwest. No one had hard data, and, most importantly, no one had maps. The true distribution of this flavor was a mystery that I needed to solve.

And so, I’ve spent a large chunk of the last couple of weeks doing so. Let’s start with the maps, and then later I’ll walk you through how I got the data. For fun, I mapped the data four different ways, so pick your favorite (and click on any to look at a larger version):

I think these maps show pretty clearly that blue moon is a regional flavor, though there are a few possible outposts elsewhere in the United States. However, they only sort of validate all the anecdotal articles about it being a “Midwestern” flavor. Minnesota is as Midwestern a state as they come, but it looks like blue moon is scarce in large parts of the state. Likewise Ohio and Iowa. The Blue Moon Core Area (BMCA), as I’m going to start calling it from now on, is really centered on a subset of the Midwest: mainly Wisconsin and Michigan, and parts of Indiana and Illinois.

These maps also explain why I didn’t realize that this wasn’t a popular flavor everywhere: I’ve lived pretty much my whole life in the BMCA.

Notice that while these maps show the distribution of blue moon, two of them also symbolize how much you should trust the map. What’s that about? Well, let’s talk about how these were built, and all the caveats that entails.

The Journey

Originally, I was hoping for some way to parse through a whole bunch of menus of ice cream parlors, searching for the term “blue moon.” There are large databases of restaurant menus out there, and there are APIs that let you access those data. But after looking through the data offerings of places like Zomato, Foursquare, and others, I couldn’t find a way to get quite what I wanted (at least for free). I considered doing some manual searches, instead — just picking a few example cities and looking at a few dozen menus of flavors, but that quickly bogged down, too.

Eventually, I gave up on trying to search for the occurence of blue moon on actual menus and decided instead to do keyword searches via Yelp. Through their API (and their website, for that matter), you can specify a location, and a search term. So, if I picked an area and searched for “blue moon,” I might get places that have blue moon ice cream. Might. And here’s where we come to the noise in the data, of which there are two kinds.

False Positives

When you search for “blue moon” on Yelp, one of the big places it looks is the text of user reviews. So, a place might turn up on my search because someone’s review said, “I like the blue moon flavor here best!” That’s great — that confirms that this place has blue moon. But, there are other ways a place might turn up on the search even if it doesn’t have blue moon. I spot-checked the data and found that sometimes a place would turn up because someone said something like “we only go here once in a blue moon.” Or one person who said, “if the prices were any higher we would just go to Blue Moon Bar.”

There’s also a brand of beer called Blue Moon, and there may be establishments that have Blue Moon in their name. Those are other sources of false positive signals.

For some other results, I noticed there were no reviews mentioning the term “blue moon,” but Yelp gave a result anyway. Not sure what they’re relying on — maybe menu information (though in the two examples I checked, Yelp didn’t offer a menu for that place)? So, here, I can’t really tell if that’s helping either the signal or the noise.

False Negatives

I have no doubt that there are ice cream places out there that offer blue moon, but that didn’t turn up on my Yelp searches. Maybe no one said anything about it in a review, or it’s not on a menu or website that Yelp could check. I would guess, though, that this sort of problem isn’t likely to be concentrated in any given area of the map. An ice cream place in Wisconsin, is not more or less likely to get overlooked in this way than a place in Arizona. So this shouldn’t affect the regional pattern we see.

And there are also, likewise, ice cream places (whether or not they have blue moon) that just don’t show up on Yelp, or weren’t correctly placed into their “ice cream” category (which I used to filter my results).

Mass Search

So, now that we have some caveats for our search data, let’s talk about how those data were collected. I did not just hang out on the Yelp website searching over and over for “blue moon” in different locations. Instead, I had a script do that for me.

It took many hours of false starts and revisions, but here’s the method I eventually came up with (on the sixth try). I created a grid of points that covered the US at an interval of 25km (in my projected coordinate system). Then I later added one for parts of Canada (I was curious if blue moon had crossed over the border from Michigan, but it turns out it largely has not). There were about 19,000 points in total.

I then had my script feed each of these coordinates, one at a time, into the Yelp search API. For the first pass, I searched for any place that was in their “ice cream” category, within 20km of the location I was searching (though Yelp warns that when you set a search radius, they take it only as a “suggestion” and will expand or shrink it without telling you, when they deem fit). The grid points are spaced at 25km, so this creates some overlap between them. The goal here was to basically find every single ice cream place that Yelp knew about. I got a lot of duplicates, which helped assure me that I was covering the area pretty well, and when I filtered through them all I came up with about 39,000 unique ice cream parlors.

I then repeated the search, but this time I included the term “blue moon” (in quotes, to keep the two words together), and I widened the search radius to 40km (the maximum) to really make sure I didn’t miss anything. This time, I came up with 771 unique places which were a subset of the total ice cream parlor dataset.

I lied: I actually did the searches in the opposite order, but I thought it would be easier to explain it if I talked about the simpler search first.



Now we know where (some of) the blue moon places are, and we also know where (most of) the ice cream places area — so we have something to normalize our data with. From here, it was just a matter of visualizing everything, to see what proportion of ice cream places also had blue moon. I first tried the ever-trendy method of hexagonal binning.

Note that there are no numbers on this map. Instead, your chances of finding blue moon just range from “not great” to “pretty fair.” Given the weakness of our data, it’s important not to say too much. Vagueness can be good, and I’m leaning here on what I have called salutary obfuscation. Giving numbers might lead people to place more trust in the map than is warranted.

Also the numbers might seem unimpressive. The largest hexagons represent areas where about 10% of ice cream places had blue moon. That doesn’t seem like a lot for an “iconic” flavor, but again, we’re likely missing a fair number of blue moon occurrences here. That’s also the reason I still showed hexagons for areas where I did not find blue moon. I don’t want to suggest that it’s actually absent from that area, because we cannot be that confident in our search. Instead I just say that the chances are “not great,” which is very different than labeling it “0%” or saying it’s not possible.

But what’s that bit about trusting the data? Well, see that giant hexagon in the southwest, on the border between New Mexico and Arizona? 9% of the ice cream places in the area came up on the blue moon search. So it looks like it’s fairly popular (given the caveat that 10% is pretty high in our data set). Except that if you look at the data, there were only 11 ice cream places in the whole hexagon, 1 of which had blue moon. So the small sample size means it’s hard to trust that this is real and not a blip due to noise. On the other hand, some of the hexagons have hundreds of ice cream places, and dozens of blue moon ones, making it them more robust and unlikely to change much due to noise in the data.

So, the “should you trust the data” question is really “how many ice cream parlors are there in each hexagon?”. Here, I’ve made the hexagons more or less transparent based on that number. Any hexagon with 250 or more ice cream parlors was fully opaque (an arbitrary cutoff), down to 20% opacity for hexagons with only one ice cream parlor. Locations with zero ice cream parlors were removed.

Kernel Density

I next tried doing a raster-based map, visualizing a kernel density surface.

Here, I did a 200km radius for blue moon places, and the same for all ice cream places. Then I just took the ratio of the surfaces. I didn’t play with the kernel density parameters too much, except for the radius. A wider radius basically means the end result is blurrier, and I wanted to be vague and blurry because, again, the data can only be trusted to give a general pattern. But, this large radius also creates false impressions at times. There’s a big thumbprint in Idaho, for example. There are a few blue moon places there, but they are all concentrated around one city. However, the kernel density spreads them out into the surrounding area, which is misleading. So there was a tradeoff here (though the large hexagons sort of offer a similar tradeoff).

Here I put stripes over the map in the areas of poor data (again, arbitrarily defined; this time to be locations that had under 100 ice cream parlors). The two symbologies don’t pair quite as “cleanly” as the hexagon method. The chance of finding blue moon is shown by a light-to-dark ramp. But the data quality stripes also darken areas, making them seem like less. It’s not too bad, but unideal. Maybe. Or maybe it’s good to darken those places and make them look like less blue moon is there because we’re not confident in them.

Two Choropleths

Now, since the first two maps in the series (hexes and kernel density) have problems with low sample sizes in some regions, I decided to make a simple choropleth map of the states.

Not much to say here, except each state had enough ice cream places that I didn’t feel the need to indicate anything about data quality. The result is that we now have a simpler visualization, which is a big advantage. Of course, states are arbitrary, and culture (including food) doesn’t stick to state lines. So this map misses out on some of the subtleties of the pattern — but, it’s also probably the most easily digestible by a mass audience.

Finally, I did one last map (a month after the others). It’s another choropleth, but this time I made up my own area units. Remember the initial list of 39,000-ish ice cream places from Yelp?

Well, I decided to use those dots to define arbitrary regions, such that each region would have roughly the same number of ice cream places. I defined 128 of them, such that each region had either 291 or 292 ice cream shops. This took a long time, because I did it manually rather than figuring out a better way. Then I made another choropleth.

The units are perhaps not always sensibly drawn. But, at least each area is now on equal footing: with the same number of ice cream shops in each, we once again do not need to worry about dividing by small numbers and getting noisy data, as we did in the first two maps.

Weird & interesting projects like this take me a long while, and I’m a freelancer. So, if you derive some value/amusement from it, please consider donating to support my continued work.

Of the four maps, I think the kernel density is the prettiest, but the hexagon map is probably best in terms of showing the pattern at the level of detail that it deserves. The two choropleths are simpler, and the state choropleth is definitely going to be most familiar to an audience.

So there you have it: some hard-won evidence for the distribution of a regional ice cream flavor. Given the subject matter, I’m pretty sure the ratio between “amount of time and effort I devoted to this” vs. “amount people care about it” is going to be pretty unfavorable compared to a lot of my other projects, but I will be able to rest easy at night knowing that I have done what I needed to.

Musings on Approximate Labels

When I teach map labeling, I explain how one of the fundamental principles is to create a clear visual relationship between the label, and the thing it’s labeling. We always want it to be plain to our audience which label is meant to connect to which dot, polygon, etc. There are a lot of strategies for this, and I talk about some of them in this video.

So here, for example, are labels that clearly go with certain map symbols, which is made clear by the position of each label on the screen, its path, and its color:

Sometimes, though, labels don’t clearly attach to anything at all, and instead float around with no particular shapes or colors to relate themselves to. As an example, take the Sahara Desert and Atlas Mountains labels here:

I love this use of labels. It lets you show approximately where something is, without needing to be too specific. It offers what I might call a salutary ambiguity. There are times when it’s good to be vague and approximate. There are plenty of things we might want to map, but which don’t have clearly defined edges or locations. What are the boundaries of the Atlas Mountains? There isn’t some hard line that demarcates the end of the highlands and the beginning of flatter terrain. And what of the Sahara? The desert does not abruptly end and turn green. So many geographic boundaries are fuzzy. Or even unknown: when you’re mapping ancient cultures, for example, it may be that archaeologists only have a rough idea of the lands that a prehistoric civilization inhabited.

It can be hard to symbolize an area that has this sort of ambiguity or uncertainty, and using a label, with nothing else, is sometimes an excellent solution. There are others, of course — fuzzy, simplified polygons come to mind. But sometimes even that is too specific. When it’s impossible or at least unimportant to show specifically where the edges of an area are, a label is your friend.

What’s interesting to me is how this highlights two subtly-yet-very different uses of map labeling. In the first example above (with the point/line/polygon labels), the labels are a supplement to the map symbols. But with our North Africa example, the label itself becomes the map symbol.

Map symbols are positioned in a way that they indicate where something is on the Earth (or whatever other realm we’re mapping). Most of the time, though, labels sort of aren’t. If you’re labeling a city dot, the position of that dot is directly related to the actual latitude/longitude of that city on the Earth. The label, though, is not positioned to indicate a location on the Earth, but instead is placed in a way that simply ties it, visually, to the dot. The city dot is basically fixed by geographic reality (though we do occasionally nudge them a bit), but there are several places the label can go. It is, at best, indirectly tied to geographic reality. It floats, tethered, to the map symbol, which is tethered to the reality.

In essence, there’s a sort of geographic layer to our map, with all the points/lines/polygons/relief/etc. indicating where stuff is. In this layer, position is meant to communicate geographic reality. And then “above” that there’s a layout layer full of other clarifying information, with positions that are not as strictly tied to geographic reality. Instead, the position of items on this layer is a bit more meta: they are referenced to the position of symbols on the geographic layer. This layer includes things like labels, annotations, scales, legend, etc.

So, when we instead use the label as a map symbol itself, to show, with appropriate and valuable ambiguity, where something approximately is, then the label now moves down into the geographic layer.

These are still ideas in progress, and I invite others to come along and refine them (or maybe someone already did so and I’m just unaware). I’m sort of making it up as I go. In truth, I started this blog post mainly because I wanted to just say, “it’s fun how labels can let you get away with indicating approximately-but-not-too-exactly where something is.” But then I started thinking about all this other stuff, and so there you have it.

The Pieces of Maps Behind Me

Sometimes people ask about what’s on the wall behind me when I’m presenting via my webcam. So, I thought I’d make a quick explanatory post.

Nailed to the wall of my computer alcove is a set of four aluminum plates which were used to print part of a book — specifically, the back side of signature 3 of the fourth volume of the Atlas of Design.

Now, if any of that didn’t make sense, that’s alright. I have them up there as a conversation piece, and when people used to visit my apartment (in The Before Time), I would use these plates as visual examples to offer an impromptu lesson on how books are made.

So, how were these plates used? Well, most color printing is done with only four colors of ink: cyan, magenta, yellow, and black (CMYK). By combining small dots of these inks in the right ratios, you can produce full-color images via a process called halftoning. Here’s how it looks at the micro-level (taken from the linked Wikipedia article):

And here’s how it looks at the macro-level, in which layers of varying amounts of cyan, magenta, yellow, and black make up the full-color image:

A book like the Atlas of Design is created using a method called offset printing. You can check Wikipedia for more details, but, the simplified version is: a big printing press will stamp each of those four colors of ink, one at a time, onto a sheet of paper. There are four CMYK inks, and I happen to have four plates. Each plate controlled the distribution of one color of ink when the book was being printed. Don’t be fooled by the fact that they all have a blue color. That’s just the color of a special coating on the aluminum, which controls what parts can pick up ink and what parts can’t. If we zoom in, we can see that they’re labeled, so that the press operator knows which part of the machine to load them into.

So why are these plates so big, when the Atlas of Design is a book that’s only about 9 by 12 inches? If you look carefully, you’ll see we’re actually printing 8 pages of content at a time.

This makes book printing much more efficient. We print 8 pages on a single large piece of paper. And then, we flip that piece of paper over, and print 8 more pages on the back side. So, each piece of paper has 16 pages. That’s for this particular book — if you have different page sizes or a different size piece of paper in the machine, the number of pages could easily be different for different books.

After it’s printed, it gets folded and cut up into the individual pages, and then it gets sewn into the book — some other books use different types of bindings, but this is how it works for the Atlas of Design (which uses a Smyth-sewn binding). This group of 16 pages is called a signature, and if you look at the spine of the finished book, you can see how it’s made of these blocks of pages, each of which started out as a single sheet of paper.

So, these particular plates also have a label indicating which part of the book they go to:

Sig: 3Back means that they’re for the back side of the third signature. I picked this particular signature & side because it features a map that I made, which originally appeared in the Ecological Atlas of the Bering, Chukchi, and Beaufort Seas. The printing company was kind enough to offer me all the plates from the book (that would have been several dozen), but I figured that might be excessive.

When I worked as an editor on the first two volumes of the Atlas of Design, I learned how this method of printing/binding in signatures impacts the book’s production cost. Since we were printing 16 pages per signature, it meant it would be best if our book length was a multiple of 16 pages. A 96-page book is printed from 6 sheets. So is a 94-page book, except parts of one sheet are just blank and eventually get discarded. But you still have to pay for them. So, the cost to print a 94-page book was basically the same as a 96-page book, since they used the same number of sheets. But if we went up to 98 pages, then we would need a 7th sheet (and the corresponding plates), and those extra two pages would suddenly add a lot to the cost.

Not all books work this way, of course; there are different printing sizes, and how the book is bound has an impact, too. This is all sort of simplified, but hopefully you may begin to get an idea of some of the interesting complexities. I really like these plates because they prompt all that thinking about the bookmaking process.

I also think it’s fun to look at how the colors are built up from these four inks. Take, for example, Jonah Adkins’s map of the One City Marathon. He has a bright magenta line running through a green background.


Now, have a look at the cyan and magenta plates that controlled how it was printed in the Atlas of Design (note that these plates only show part of the map; the rest of the map was printed on a different sheet of paper):

On the cyan plate (which, along with yellow, will create a green background), there’s a hole for where the magenta line will go. On the magenta plate, most of the background is absent, leaving just the route line.

These plates are full of all kinds of these fun details, where you can see the artwork being built up piece by piece. Even now I still stop to look at them every once in a while.

So, that’s what’s behind me when you see me on the internet. You’ll mostly just see pieces of the cyan and yellow plates, but there’s a lot more neat stuff going on, all of which was needed to print just one part of just one book.

Back to the Rivers

Friends, a few months ago I finally published An Atlas of North American Rivers, a series of maps showing the connectivity of major stream systems across the continent, done up in a style reminiscent of transit maps. It was a project that I’d left fallow for many years and finally, due to the pandemic lockdown, finally wrapped up after nearly a decade of letting it sit on the shelf.

I was very relieved to have it off my plate after so many years. But, there was one loose thread dangling: I’d long intended to also put together a final piece that combined a lot of the atlas’s content into one single poster. So, a few weeks ago, I finally got around to doing just that.

Feel free to click that image to see a larger version. There are over 800 labels, so there’s plenty of browsing to be done.

I’m also pleased to announce that, if you like this map, the Transit Maps store is now offering prints! Most prints of my work presently sell via Zazzle, one of the big print-on-demand companies. But I’m excited to partner with Cameron Booth, who runs the Transit Maps store (which has plenty of other neat stuff you should look at), to make this piece available.

So, if the above photos entice you, head on over and grab a high-quality copy!

This map represents a major revision of the design language I used in my original atlas. As I said in that blog post few months ago, while talking about how I’d established the style when I was a more novice designer, “I would quite likely do the whole project differently now (though I’m not sure how).” Well, I figured out some of the “how.”

  • I’ve changed the typeface over to Mostra Nuova, which you’ve probably seen in a lot of my work (because I’m going to squeeze every bit of use I can out of that $80).
  • The river names now match the river colors, and are angled to follow the lines.
  • I’ve given a subtle inner glow to the states/provinces, to help separate them better.
  • I’ve used different color schemes for different countries.
  • The colors have been “rationalized.” The various greens of the US states, for example, are now simply tints of the same base color, instead of four ad-hoc creations. The river colors also now exist on an even gradient between two colors, and now use only 3 inks instead of 4.
  • International boundaries are now distinguished with a white line.
  • I have been much more willing to simplify the geometry of the rivers (and therefore distort the underlying states/provinces).

For reference, here’s a snippet from the old atlas.

Some of these changes I’d pondered years ago, but was oddly resistant to, like angling the river names or adding glows to the states. But now I’m glad I made the changes. I think the end result looks a lot better.

I decided to keep the map somewhat more sparse than some of the atlas pages. For the atlas, I was often trying to capture every single city and village along the route. Which made it pretty busy at times:

For this poster, I’ve spread things out a bit more. I think there’s a fair density of settlements, but everything still has breathing space.

Finally, I did think about making the poster of all of North America, but there would have been much more ocean on that map, and far fewer rivers. I decided to go with the continental US, plus a fair chunk of Canada and Mexico, which allowed me to fill the space pretty well. Perhaps, another time, there will be another one covering more of the continent. Or perhaps someone else out there will do it.

Meanwhile, I hope you enjoy this (final?) piece in my river maps saga. And if you’d like to put one on your wall, head on over to the Transit Maps store!


I used to run another map blog, before this one.

It’s an effort that I look back upon with regret, and prefer not to think too much about. After leaving it fallow for years, I finally took it offline a year ago and hoped nobody would notice, or remember it even existed. However, I think it’s time that I more publicly own my mistakes.

Cartastrophe was a map critique blog, in which I took other people’s maps and pointed out their flaws. There was a lot of sarcasm. I’ll spare you any quotes because I think you know how it goes; there’s plenty of similar content out there in social media right now.

I started Cartastrophe because complaining about the work of others was easier. See, I’d originally planned to run a blog like the one you’re looking at now — discussing my designs and my thoughts on cartographic processes. But, as a post on Cartastrophe recalls:

it quickly turned out that I didn’t have much to say on the subject. So, instead, I closed [somethingaboutmaps] down and started Cartastrophe, because I had plenty to say about other people’s maps.

Apparently, just not blogging wasn’t an option for me. Fortunately, I eventually found that I had actual constructive thoughts to share that didn’t involve criticizing other mappers, and so I resurrected somethingaboutmaps and posted less and less on Cartastrophe; it was mostly quiet by the end of 2012.

It’s worth mentioning that I did try to make Cartastrophe more than a place for simply complaining that some mapper had done a bad job — I wanted to use these examples to teach. As with any good critique, I tried to explain my rationale: why I thought certain things should be changed, and what this person’s “mistake” could teach us about design and human perception. I also required myself to say a minimum of one nice thing about each map, and I occasionally posted analysis/critique of maps that I really liked. I learned, however, that for me, it is a lot harder to clearly express what I like about something than what I don’t. Finally, I tried to show that I and others were not immune to mistakes: A few colleagues and I posted critiques of our own work.

In the end, though, most of the site was me posting what I thought were “bad maps,” and telling people how I thought they should have been done better.

I took people’s maps, uninvited, and publicly stamped my thoughts on them. I did not ask the authors about their goals or process; I made assumptions, instead. I did not ask them if they were comfortable with a public critique. I did not ask them what they thought about the work — maybe they didn’t even like it (my maps sometimes feature parts I don’t want to claim credit for, as clients push me to make decisions I disagree with). I did not invite them to be a part of the process of improvement and learning. They never had a chance to explain themselves before I passed judgment.

Now, I’m not suggesting there’s absolutely no value in looking at other people’s designs and trying to learn what we might want to avoid, nor do I suggest we stop having negative thoughts about the works of others. But it’s all about the approach and context: my good and/or educational intentions did not matter as much as the importance of including the original map author as a partner in public critique, which I rarely did.

Now, someone’s going ask, “should we never publicly call out a grossly misleading map without the author’s permission?” That’s not what I’m saying, and that’s not what Cartastrophe did. I wasn’t looking at maps that were serious threats to public knowledge and warning readers about them. I was nitpicking the design choices of innocuous maps that were perhaps confusing or difficult to read. It’s one thing to say “The public must know that this particular map is incorrect about something important,” and quite another to say “this map about tectonic plates has an illogical color scheme.”

I ran Cartastrophe because it was an easy way to get attention when I was in graduate school. It was easier for me to point out flaws than cogently praise excellence, and it was easier to write quips about the failings of other people than to form coherent thoughts about my own cartographic practice. And it was easier to feel I was a good designer if I could break down ways that other people were not. That’s the core of it. I am certain I’m not the only one to offer a critique for such reasons. I, and probably some others, learned a lot through the process (again, I won’t deny that this stuff had pedagogical value), but my approach was rarely one of partnership, and instead one of “I’m the expert and I’m here to school you (but I’m making jokes so it’s all OK, right?).”

Also, I don’t even like puns. I have no idea why I liked the name “Cartastrophe.”

I hid the site last year, without fanfare. I didn’t really want to draw attention to my mistakes — perhaps I wanted simply to avoid the public criticism that I had given others. I was content to bury that whole embarrassing business. But today, I saw an excellent and quick talk by Amber Bosse, in which she discussed cartographic gatekeeping. I remembered that, in Cartastrophe, I had once been a full-throated gatekeeper. And so I thought it was time to say that I should have done better; I likely hurt people, and I am sorry to have done so. I will surely continue to make mistakes (hopefully different ones!), and I can only hope that my colleagues will be kind enough to continue showing ways that I can be more positive in my contributions.

Addendum: I thought of one more point, about a day after I posted this.

Lessons Learned: How to do Map Stuff

Friends, I was very pleased (and overwhelmed) to have hundreds(!) of you join us during the How to do Map Stuff event yesterday. If you missed some or all of it, no worries! Most everything was recorded. You can find most links in this YouTube playlist.

There is one other recordings that isn’t on that list yet:

I had never done anything like this before, and neither had most of the presenters, so here, I want to share some loosely organized reflections and lessons learned.

This whole thing was pretty ad hoc. It came from an idea I had in the middle of the night: I already do occasional map livestreams, so maybe if a group of other map people did them on the same day, that would be fun? I announced it without much ahead planning, and made things up as I went along.

Things quickly grew beyond my expectations, Originally, I was worried that I wouldn’t be able to get more than one or two other people to present, and that we wouldn’t have more than a dozen people in the audience (which I would be fine with, but I worried would make the other streamers feel bad). But I decided to give it a go, nonetheless. We ended up with audiences in the hundreds, and 26 streamers, which was amazing, but also offered some challenges, as you’ll see below.

Time Zones

When you’re planning an in-person conference, you can safely assume that your presenters and attendees are all in the same time zone. But How to do Map Stuff presenters came from around the world, from Hong Kong to Louisiana to Munich, and so they were all awake at different times. When it came time to arrange the schedule, I had to scramble around on Twitter and other sources to figure out where people lived, so that I didn’t ask them to present during a time when they should be sleeping. When presenters signed up for the event, I didn’t think to ask them what time zone they were in, but I would certainly do so now.

During an in-person conference, your presenters are usually available for the entire duration of the conference; they’ve taken time off from work to be there. But, most people aren’t going to take the day off for your virtual event, and so one other thing I didn’t plan for was that people need to fit in work alongside their presentations. Fortunately, I didn’t run into too much trouble there, but I did make a couple of rearrangements to fit people’s work schedules.

In sum, my advice to you is: when presenters sign up, ask them what time zone they are in, and when they are available. Having that information in advance would have allowed me to schedule everyone with much less hassle (both for me and for them).

The distribution of my presenters pushed me to think about the time zone of the event’s audience, as well. I mostly know American cartographers, and so I figured that they would be the only people who might hear about it (again, I didn’t expect it to generate so much enthusiasm). So, the schedule is centered on time zones in the US/Canada. But, I made the event longer than I had originally planned, pushing the schedule to start earlier and end later. In this way, folks in Europe/Africa would be able to catch a few hours of presentations before going to bed, and people in East Asia and the Pacific could wake up and see some of the event as well. I think the long schedule may have contributed to some audience fatigue, but I’ll talk more about that later.

International Date Line

This is really an extension of my comments about time zones, but it’s important enough that it gets its own header. Due to the International Date Line, your event may be on a different day depending on where people live. I live in Wisconsin, where How to do Map Stuff was on April 29th. For a person in South Korea, though, it was on April 30th. I tried to be very clear, on the schedule, on tweets, etc., to indicate both days, so that no one got confused.

To keep everything fairly clear, with an audience scattered around the world, I listed six time zones for each presentation in the schedule, as well as some date indicators, so that people would have some reference points.

Screen Shot 2020-04-30 at 10.40.00 AM.png

Google Sheets

Speaking of scheduling, I made an ad hoc decision to publish the schedule on Google Sheets, which turned out to be a bit of a problem.

During the initial sign-up phase, someone on Twitter asked if I could post the list of talks that had already been submitted. So, I linked a Google Sheet into the Google Form I was using to accept submissions. And people started circulating the Google Sheet link, and it became one of the main places people learned about the event. So, I decided to just formalize it and make that link the go-to place for the schedule.

However, it turns out that only 100 people can look at a Google Sheet at once before it starts to lock people out. And at times, more than 100 people wanted to see the schedule. Fortunately, someone on Twitter mentioned that they could not access the schedule, and I looked into it. It turns out you can hit a button to “Publish” a Google Sheet, which makes it so that more people can view it, while reducing some of the features (no one was going to edit it except me, so this didn’t matter). But it also moved the schedule to a new URL, so I had to scramble to circulate the new link one day before the event. Also, the “Published” version of the sheet can take up to 5 minutes to display updates that I might make, which can be a bit of trouble when making last-second updates to people’s video links.


It’s no particular secret that, when looking at conference speaker lineups, women are underrepresented. At a large conference like NACIS, women typically deliver 35–40% of the presentations. About a third of the presenters at How to do Map Stuff were women, which, while in line with other events, is certainly not great. At first it looked like that percentage might be even lower; as speakers began to sign up, there were almost no women among them.

I wondered why this even might attract proportionally fewer women, and my colleague Meghan Kelly suggested that it might be because women disportionately are tasked with childcare and other household management, both of which are in high demand now that schools are closed and families are home more. I also came across this tweet thread which points out how the same phenomenon is affecting journal submissions.

Now, I mentioned that, while things started slow, How to do Map Stuff eventually reached about the same percentage of women speakers as a conference like NACIS. I think this is partly because I tweeted out an appeal asking specifically for more women to sign up, and the folks at the Women in GIS and Women in Geospatial+ Twitter accounts, among others, were kind enough to help spread the word among their contacts. But, in hindsight, I’m not sure about this approach. If women are volunteering to present because they’re much busier than usual during quarantine, is it fair to then make a special appeal to them to do more work by presenting at an event?

Once I had my speaker list in hand, I tried to amplify (or maybe not de-amplify) women’s voices a bit in the schedule, by making sure they were never scheduled to go at the same time on adjacent tracks. I’m not sure if it helped at all, but I thought it was worth a try.


I initially only built one break in the schedule, about halfway through. A few others eventually appeared due to presenters dropping or shuffling around. Since all of the audience was operating on different sleep/eating schedules around the world, I could not really plan a “good” time for breaks. So I left it to them to wander off and come back, as needed.

However, more breaks would have been good to have in the schedule, because they offer slots for on-the-fly rescheduling. Due to technical problems, one presenter could not make their original time slot (and some others had close calls). But, that presenter eventually figured those problems out and was able to move to a break later in the day, with enough time to spread the word about a schedule update. We could also have just had a third track, but I wanted to avoid splitting the audience too much, and so it was nice to have a space in the existing track for these sorts of emergencies.

Even though they might not have been ideally timed for the food/sleep/bathroom/email needs of every audience member, I expect viewers might have welcomed another break or two in the schedule, as well. They would have offered a bit more time to catch up, digest, and ponder.


All presentations were back-to-back, with no gaps. So, viewers needed to immediately switch over to another presenter. I’m sure some viewers started tuning out when a presenter was nearly, but not quite, finished, as they prepared to move to the next person. None of this was ideal, though it wasn’t a huge problem.

I went with this schedule because there’s an easy-to-remember appeal of having every presenter start on the hour, or the half-hour. Television shows keep this kind of back-to-back schedules (though they have commercials for padding). I figured it would be more cumbersome to tell people, “This person starts at 1:05, and this other person begins at 3:20, etc.”

Probably in the future I’d just ask presenters to conclude 5 minutes early. This keeps the even scheduling of 30- and 60-minute time slots, but still gives audiences time to refocus and move on to the next person.


As in all conferences, some people ran out of time, or ran over their time. One nice thing about doing all this online, with separate video feeds from each presenter: if one person exceeds their time, they don’t prevent the next person on the schedule from starting on time. The audience can switch over if they choose. Of course, it also means the two speakers are competing for audience, which isn’t ideal, but at least it’s not quite as bad a situation as when this happens during an in-person conference, where the audience has no choice and the subsequent speaker just has to wait. I wasn’t able to catch the end of everyone’s presentations, but I did hear a few people pushing themselves to wrap up on time so that the audience could move on, which I appreciate.

Zoom vs. YouTube

Let’s get to the big one. Most people who’ve done screensharing and web presentations have done so through Zoom, or Skype, or a similar video-conferencing program. I’ve also seen some in-person conferences select Zoom when switching over to an online conference. Zoom is familiar to a lot of people. But instead, I encouraged presenters to do an entirely different workflow: livestreaming on YouTube. There are some good reasons for this, as well as some disadvantages.

To a certain extent, it’s about choosing the right tool for the job. Zoom/Skype/etc. are designed specifically for peer-to-peer conferences. For meetings between you and your friends/colleagues/etc. where more than one person is expected to speak, and where interaction needs to be close to real-time.

Professional (or at least frequent) livestreamers (of games, crafts, etc.) rarely use those programs. Instead, they use special software (like a program called OBS) to capture video of their desktop, and then send that video feed to YouTube/Twitch/etc., where it’s distributed to an audience via a website (vs. Zoom, where you need to run an external application). This workflow is designed for presentation from a single person, rather than Zoom’s peer-to-peer interaction.

Zoom is familiar to a lot of people, and it’s fair to say it’s easier for presenters. I think the audience experience isn’t as smooth, though. Audiences have to download Zoom and leave their browser, instead of skipping from one simple YouTube URL to another. The chat isn’t as clearly integrated into the presentation; on the Zoom presentation I watched, I saw less chat interaction from the audience than I did on YouTube streams. These bits of interface may seem like tiny thing, but an extra click or two really matters in terms of audience experience and engagement, as UI designers will tell you. YouTube makes it just a hair easier on audiences, and I think that’s important.

YouTube also offers one nice archiving advantage as well: once your stream is done, it stores the video recording of your presentation at the exact same link. So anyone who misses the presentation doesn’t need to find another link to watch the replay. Another small way the audience experience is smoother.

Streaming via livestream software also allows for fancier setups than just a simple screenshare. As we saw with some presenters, they overlaid their camera feed onto their desktop, or switched to static images, or even had cool animated transitions between scenes. So, there’s more opportunities to get creative, or make a more engaging audience experience.

There are no audience limits on YouTube. Zoom only supports 100 people for free, and we had a couple presentations were people couldn’t get in. Of course, that was also because we didn’t realize that so many people would want to show up to the event (I clearly remember telling one presenter, weeks ago, that the audience limit would not be a problem).

Zoom allows audience members to ask questions via audio (and video, if you like), which is more like in-person conferences, and that is nice to have at times. YouTube will only let people type in the chat window. But on the other hand, a chat window lets people ask questions without interrupting the presenter (and lets audience members answer each others questions), which is a different sort of advantage. Zoom has a chat window, too, but again, it’s not utilized as much. Also, since Zoom defaults to letting the whole audience have audio, you also have issues where someone’s unmuted. During one person’s presentation, an audience member had music going on in the background while they were unmuted, and so the presenter had to ask them to mute themselves (or track them down in the list of dozens of participants and mute them).

So, in sum, I’d say: Zoom is generally easier on presenters, but a bit less smooth for the audience, and streaming via YouTube and broadcasting software is harder on presenters, but gives a nicer audience experience. They each have their upsides and downsides (some of which I haven’t covered here), but having presented on both of them, and been an audience member on both of them, I’d much rather use YouTube for something like How to do Map Stuff. Zoom et al. still have their place (and I use them for other things).

So, that’s why I asked presenters to go the YouTube route (also, it’s because that’s what I was familiar with, and I was only familiar with it because when I started streaming last year, I just looked at what other people were doing, and it seemed to be the way to go). And I must emphasize that the presenters did a lot of work to make that happen. Most of them had never livestreamed before, and they took the time to learn new, more complex software, and work through technical problems.

Technical Problems

And there were technical problems. There was the online equivalent of “my laptop doesn’t connect to the projector,” which was people’s feeds not quite starting. One person’s computer crashed in the middle of the presentation.

What was encouraging, though, was to see the support from the audience during these times. There were various messages of “we believe in you!” or “we’ll wait for you!” as people worked through issues. We are fortunate to have a supportive community, and that’s something that makes me feel more comfortable when I do my livestreams. Because things will certainly go wrong when you have this many people presenting.

And sometimes there’s not much you can do about it. But I emphasized to the presenters that if all else fails, just tell the audience to go get a sandwich and enjoy the next set of presentations. Because, in the end, we’re all just casually coming together for free presentations, with no pressure. I think the audience comments demonstrated that. Also I will now use the phrase, “sandwich time” to describe a problem where you have to give up on what you were planning to do.

Some presenters had problems with video/audio quality, which surprised me. One person mentioned that they had shared their screen very often through other applications and hadn’t had any issues, but when they streamed to YouTube, suddenly their framerate was poor. I’d not run into that, but several people did. I couldn’t think of a reason for streaming to YouTube to be worse than doing a Zoom call, but I ran this by someone who’s been livestreaming for years, and she said:

One really needs to actually dig into the stream settings and tweak the max bitrates to accommodate their connection, but that’s definitely a steep learning curve for a first timer and requires multiple streams to dial in.

So, it sounds to me like maybe Zoom/Skype/etc. are more generous/forgiving in their settings, while OBS (the broadcasting software people were using to stream to YouTube) requires a little more fine-tuning to get right. It’s something I never thought about, and thus didn’t prepare presenters for, because my streams have always gone fine with the default settings OBS had suggested for me.


One thing that presenters and audience members noticed was the latency: the time delay between when the presenter said something, and when the audience heard it. That’s not actually a bug, but a feature. The latency seems to be a little larger nowadays than it used to be (I saw 15 seconds on my stream, vs. about half that last year), but in any case, it’s meant to buffer the video for the audience. We’ve all been on Zoom/Skype/etc. calls where the video gets fuzzy or the audio cuts out. Sometimes that’s due to issues with the person sending (see above about fine-tuning OBS settings), but sometimes that’s also due to your internet connection. If you remember the early days of internet video, they would sometimes pause and stutter as your incoming data stream caught up. So now, videos are buffered, and YouTube handles this for you by introducing a delay between the presenter and audience, to prevent this particular form of quality reduction.

It can sometimes be a bit awkward to communicate with your audience on delay, but I’ve found you (and they) will get used to it quickly.

Audience Fatigue?

I noticed, as the day went on, fewer people were tuning in. We had 300–400 early on, and by the end of the event, we had about 75. Some of this is likely due to time zones. Europeans were asleep by this point, and as evening approached for Americans, many of them probably switched off to focus on dinner and family time. Folks in East Asia & the Pacific started tuning in, but I think there were proportionally fewer of them, especially as it was a conference broadcast in English.

Some of it, though, could simply be audience fatigue. It may have been fun and enlightening, but it’s still a lot of content to absorb, and our brains become full and I expect some people just got tired and decided to catch the recordings later on.

So, instead, it could have been good to split this into multiple days. However, I would want to change around the hours of the presentations each day, so that there was still content easily accessible (time-wise) to folks outside of the US. So, maybe one day that starts early in the United States (for people in Europe/Africa to watch), and one day that goes later (for people in East Asia/Pacific).

Again, I was just hoping to get 3 or 4 presenters, so I never planned to accomodate this idea. Once I had to schedule over 25 presentesr, I did consider expanding the event, but I figured it was too late. I had already advertised the day, and presenters were planning on presenting on that particular day, and I didn’t want to ask a bunch of them to move. It would be something to plan in advance, next time.


I’ve seen some people on Twitter saying they hope this sort of event happens again soon. That’s a nice thought, but it could be a long while before I’m up for organizing an event at this scale. This turned out to be more work than I had anticipated. I thought at first that I would just collect presenter names, put them in a schedule, and then everyone would do their own things. And, certainly, the presenters did shoulder most of the labor burden; they deserve a lot of praise for their effort.

Nonetheless, organizing was still more of a challenge than I bargained for. Before the event, there was scheduling and re-scheduling, and there was a lot of advising of presenters as many of them prepared to livestream for the first time. Once the event started, I didn’t catch most of the content — I was switching constantly between

  • monitoring Twitter,
  • answering questions on my blog and by email,
  • monitoring the start of everyone’s presentation to make sure they got started correctly,
  • helping troubleshoot when they did not,
  • making last-minute schedule adjustments and announcements,
  • watching streams for problems (rather than content), and
  • being a ball of stress and nervousness (I’m a nervous presenter, and so I was empathetically nervous for most people going live).

I do not recite this list for your praise or pity; this is mostly for the benefit of future planners of such events, so that they have a sense of what they might encounter.

So, I definitely need a break =). But, it’s important to note that such events in the future don’t have to involve me at all. My hope is that the presenters will all keep going with livestreaming on their own, and perhaps inspire some of the event’s audience to do so, as well! Even before the pandemic, I felt that livestreaming was a good fit for cartography, and we could use more of it. We can share planned demos, or even just turn on a feed for a few hours and let people drop in to watch us work.

Worth It

Despite the stresses, it was certainly worth it. I hadn’t really anticipated the response we would receive. Hundreds of people tuned in, from Egypt to Brazil to Indonesia to Sweden. That still seems crazy to me.

More important than audience sizes though, was the emotional impact. I had thought people would find How to do Map Stuff interesting, and fun, and maybe just a little bit connective. But the response was emotionally deeper than I had expected. More than one person shared with me the great joy they felt, how they were buzzing with adrenaline, and how, importantly, it lifted their spirit during a rough time. That means a lot to me to hear.

Odds and Ends, Part 2

I do a lot of mapping for fun and exploration. Sometimes these projects get their own blog posts and end up in my portfolio. But other times, they’re little things that don’t really have anywhere to live; they’re not worth blogging about on their own. Instead, I generally scatter them on the winds of Twitter, and move on to something else.

However, I want to give some of these cartographic trifles and doodles a more stable place to live, and so I’ve gathered several of them here. I browsed through the last several months of my Twitter account, and here’s what I’ve come up with.

Pink Things

I don’t get to use colors in the pink family very often, so I made a couple of maps to just enjoy that part of the palette.
Pink Thing 2-01 copy.jpg


0004 copy.jpg

You can click those images to have a look at larger versions. I simulated two-color halftones for each one, as I’ve been fairly obsessed with halftoning lately.


For the map of Alaska, and the relief, I just recolored maps I had previously made.

Cartographic Efficiency

I make a lot of maps of Michigan, and it’s a weird-shaped state that doesn’t really efficiently use a rectangular page layout. So I decided to see how it ranked vs. other states, assuming that we mapped them with north towards the top of the page (but remember #northisasocialconstruct).


I did a little scripting to just quickly put each state on a Lambert azimuthal equal-area projection, tangent to that state’s centroid. Then I looked at the minimum bounding rectangle, and compared its area to that of the state. That just gave me a table of values. Then I spent a long long time actually making the graphic. And to answer your question: that’s Colorado in position #1, with Wyoming very close behind.

The Upper Peninsula

This isn’t quite a cartographic doodle, but it doesn’t have anywhere else to really live in static form. I made it so that viewers could watch me label it in a YouTube video. In the end, the video (somewhat not-smoothly) pans around the map, but I wanted to put it here in its final form.


I halftoned it (again, a current obsession). Only yellow, cyan, and black inks here; no magenta. I used my giant Michigan landforms map as a reference to know where to put some labels. In the context of this map by itself, some of them probably don’t make sense, because the coarse hypsometric tinting just doesn’t make them apparent.

My Apartment

Early in the COVID-19 lockdown, people were making joke maps about their daily commute, showing things like transit maps routing people from bedroom to bathroom to home office. I didn’t make one of those, but they reminded me that I’d never seen an actual floor plan of my apartment. So, I made one.


Ice Hike

Every winter I lead folks from the UW Cartography Lab on a short hike across a small part of Lake Mendota, and I made a little flyer promoting it.


This year there were two of us hiking, which is not unusual, but I have had groups of six or so.

Map Layers

In most people’s workflows, maps are made from layers. I just wanted to do a little animation that showed how that worked. It was a good way to learn some new stuff in Blender, and get more animation skills under my belt. I’m still quite the novice, but this one turned out nicely.


So this concludes our tour of some of the random stuff I’ve put on up Twitter since mid-2019. There’s more, but I think these are the most worth collecting into one place. I’ll try to do these every once in a while. As the post title suggests, this is not my first such collection; I wrote one up many years ago. But I’ll try not to let so many years pass until the next one.

An Atlas of North American Rivers

Around a decade ago, I started making river maps, in a style reminiscent of transit networks. I made a lot of them, and had the idea of compiling them into an atlas. I wrote about them, and even sold prints of them.

And then I stopped. My atlas sat, 99% complete, mostly untouched, for about eight years. But this is a time for finally finishing things, and so I finally buckled down and wrapped up something that has lingered far too long.

I present to you An Atlas of North American Rivers.

Click on that image to download a PDF. It’s designed for print (so there are some spaces in the middle of spreads to account for the gutter), but I’ve rotated some of the pages so that you don’t have to crane your neck when I switch between landscape and portrait layout, and I’ve split up multi-map spreads. So: expect some changes in page dimensions as you scroll through.

I’m making this project free to download; if you enjoy it, you’re welcome to make a donation to support my work.

There are two reasons it took so many years to finalize this project. One is that, the longer it sat, the more I changed as a designer, and the more “outdated” these felt. I still think they’re fine, but I would quite likely do the whole project differently now (though I’m not sure how). But, I did not wish to let the perfect be the enemy of the good, and so I am letting them go into the wild, as-is.

The second reason is that, after some of the early maps went briefly viral in 2011-ish, I was confronted by many people who were unhappy with my choice to not include certain rivers or towns that were important to them. It was a stressful situation. However, I’m more comfortable ignoring those people now. In truth, though, that experience also offered a valuable lesson. Cartographers are comfortable and familiar with the generalization and abstraction that maps offer. A lot of the public isn’t, though; they don’t take it for granted the way that we do, and that’s useful to remember.

Interested in a hardcover version? I’ve got a form for you to fill out! If I get enough people to make it economically feasible, I’ll look into getting some printed.

Update: demand was extremely low, and though the book was designed for print, I am not able to offer a print option at this time; print-on-demand services were also generally too expensive to meet the cost that people said they’d be willing to pay.

However, you can still purchase prints of most of the individual maps. They vary a bit from the book layout, both in typography, and in some cases in format. For a few river systems, I designed standalone posters, but for the book I split them up.

2020 Cartographic Freelancer Survey Results

Several weeks ago, Aly Ollivierre and I posted a survey on pay and business practices in the freelance mapping community. Fifty-six of you were kind enough to take the time to answer our questions; our thanks to all of you for helping bring more transparency to freelancing!

If you’d like to see the survey results, including the questions asked, you can download the data here. The results of certain questions have been removed to ensure anonymity. If you have any questions, or find any interesting insights, feel free to contact us.

Below, we discuss the results in the same conversational format that we used to present the findings of the 2018 edition of the survey. This time around, Molly O’Halloran was kind enough to join us and bring her insights, as well. And thanks to Aly for making all the graphics!

Daniel Huffman: Ok, so, let’s start with the big number, which is: what’s the median income for freelancers right now?

Aly Ollivierre: We should first note the average, since that’s what we used last time: $65 for 2018 and $70 for 2020. Since we decided to use medians this time around:

Daniel: Right, we had one outlier rate that was raising the mean a fair bit, which would have been $79 otherwise. Median is less susceptible to outliers, so it might be a good comparison going forward in these surveys.

So, the good news is that it looks like rates are going up, whether you slice it by mean or median. And I like to think the survey in 2018 is a part of that: we asked people if their business practices were influenced by 2018’s survey, and we separately asked if they’d given themselves a raise recently. Of the 11 people who said they’d been influenced by the last survey, 10 (90%) of them also said they’d given themselves a raise. That doesn’t explicitly mean correlation, but it seems anecdotally strong. Of the 13 people who saw the last survey and said they hadn’t changed practices, only 5 (38%) gave themselves raises.

Molly O’Halloran: I was just going to say that. Information sharing, and getting each person to consider their business, is so valuable.

Daniel: Agreed! It was valuable to me personally. I raised my rates after seeing that I was often charging below average.

Aly: I also became more diligent about ensuring that I wasn’t undervaluing myself when I gave clients quotes.

Molly: Yes! Or bringing down the value of our market in general. If some people are charging rock bottom, that hurts everybody.

Daniel: It’s hard to do, I think. I always have a fear that I’ll lose business if I aim too high. You have to hit this hidden magic number. Molly, that also reminds me of what one of the people said in the survey comments:

“A rising tide lifts all ships! Everyone charge more!”

Molly: What do you do to educate the client about the value you’re delivering? Have your communications changed during the negotiating/estimating phase of potential projects?

Aly: Honestly, I’ve just found that I spend an extended period of time re-reading and stressing over the wording of the email with estimated rate/cost and try to remain firm with myself that this is what I should be charging for this work! (It’s hard.) I’ve found that being confident in my prices hasn’t resulted in anyone questioning them … usually questions instead arise from what can be done differently to fit what they’re looking for into their smaller budget.

Molly: Aly, I feel that so much. I’ve been working for years on convincing myself of the value of my work. It’s as much an internal struggle as it is a market problem, I think. This is probably an aside, but I’ve been trying out the advice I learned in a webinar and book by Emily Ruth Cohen, Brutally Honest. I used to hedge and fuss over estimates: this map will cost, say, $380. Her advice is to round up, as it projects confidence and is likely a more accurate number anyway. So that map is now estimated at $500.

Daniel: I don’t know if it’s a bad idea, but I usually hedge a bit in my emails and let them know that it can be negotiated if it’s outside their budget.

Aly: I always hedge in my first draft and then try to edit my email to find that perfect balance of showing them I’m flexible, but that I also know my worth.

Daniel: This subject brings us around to a couple of other survey results that tie in: experience and education.

Aly: I combined 2018 and 2020 results into one graphic, which I think simply demonstrates that a different crew took this survey more than anything specifically related to the market.

Daniel: In 2018, we didn’t see any correlation between someone’s rate and their experience or educational qualifications. I didn’t see any correlation this year between education and rate, either.

But, experience was different. Last time we asked how many years of experience people had, and this time we tried to get people to think more about their general experience level. Because making maps occasionally for 10 years might be equivalent to making them full-time for a year or two. So this time there was a clearer pattern:

Given the low sample sizes, I wouldn’t read too much into the dip for “expert” level, but I think we can basically say, people just starting out charge less, and then rates rise quickly and plateau.

Aly: I think this is a better way of looking at things than we did in 2018, and follows pretty well with my hourly rates throughout my freelance career (starting at $20–30 early up as an undergrad student and recent grad, raising and sticking around $40–50 for awhile, and then moving up into the $60–70 range).

Molly: Are those numbers how much one actually takes in per hour spent mapping, or the hourly rate? Maybe I’m the only one who routinely spends more time than I estimated for each map, but I reckon not.

Daniel: This is what people answered for the question “In 2019, how much money did you typically receive per hour of time spent on freelance mapmaking work?” — so, people were encouraged to think about how much time they actually spent, and what they actually earned.

Aly: I definitely want to chat more about flat rates vs hourly rates, Molly! So instead of just three categories like we did in 2018, we did 5 categories. This gives a little more of a breakdown, but honestly it’s not a big difference between the data we received in 2018 and in 2020.

Daniel: The question was turned into a score of 1–5, with 1 being hourly and 5 being flat, and I just ran a quick average and it’s 2.9, so there’s still not a dominant approach. Looks like most people do some of each. I keep trying to push myself to do more flat rate, though, because if I do a job too efficiently, I get paid less.

Aly: There are pretty strong pros and cons for doing hourly and doing flat-rate. I often find myself with clients who want a lot of back and forth which ends up eating up time and hurts me in the long run if I do flat-rate. I generally prefer hourly because then I know all of my time is paid for, but I have also started warning clients when I do flat-rate that they’re getting to the end of the finances set aside and that we will need to renegotiate the contract if they want a lot more edits.

Molly: I always do flat-rate that includes one or two rounds of revision (negotiated in the estimate phase), then bill hourly for revisions beyond that. Clients should be paying for your expertise as well as your time. It took years for you to be able to make that map so well so quickly. It’s true, though, that I often spend waaayyy more hours than I need to and then, since it’s flat rate, my effective hourly rate goes way down. Aly, your approach sounds smart.

Daniel: As one participant said:

“You’re not paying for the two hours it took me to do the map, you’re paying for the years it took me to learn how to do that in two hours.”

Aly: That point by both of you about expertise = faster mapmaking is definitely why I struggle to force myself to do more flat-rates, I just have such a hard time estimating accurately from step one!

Daniel: I am terrible at estimates. Getting better, but I sometimes bounce them off other people first to see if they make sense. And I usually am secretly thinking hourly, so my flat rate is just “I think it will take X hours at Y per hour,” plus a bit of padding in case it goes wrong.

Aly: That’s how I try to estimate my flat rates as well.

Molly: Same! Though I tend to think in half-days or days rather than hours. How many days would I be working on this?

Daniel: It’s comforting to know that we all use similar methods. It’s one of those little things that you make up and never seems worth asking about, even though you wonder if you’re doing it the best way. But that’s what the survey, and this conversation, is for.

Aly: I feel like this topic rolls well into whether or not people use contracts.

Daniel: Like the hourly vs. flat rate question, we changed this to a 1–5 scale, but in the end saw roughly the same result when compared to 2018, it looks like. The average of everyone’s results was a score of 2.75, with 1 being “never have a contract” and 5 being “always.” So, people are leaning a little toward no contract, but they’re still pretty common.

Molly: Yeah, pretty even split, too.

Aly: I rarely use a contract (usually only when my client provides it), but I feel like I really should and just don’t know where to start with it! Especially since it adds a level of intensity to conversations with clients that I usually don’t have.

Daniel: Almost all my contracts are provided by clients, on big projects. For a lot of my work, it’s some small one-off thing (a typical example is a retired history professor who wants a quick greyscale map for a book they’re writing), and it feels like overkill to write up a contract for it. But in truth, it’s probably a good idea. Some people have boilerplate ones ready to go.

Aly: Exactly, I mostly work with smaller scale clients as well where it feels like overkill. I’d love it if there was a common, simple, contract template that cartographers generally used/built upon as necessary. One respondent noted:

“I recommend flat-rate contracts with X free revisions, subject to per-revision fee of Y after the cap’s met. Add strict payment dates and a deposit of at least 25%.”

Which really seems like something good to keep in mind!

Molly: I don’t think that was me … but it could’ve been. Agree completely with the above.

Daniel: My guess was going to be that people who are more full-time might be more likely to have a contract prepared and use it a lot, while people who map occasionally would let these things slide and be more ad hoc, but it looks like that’s not the case. I grouped people by their answer to the contract question, and averaged what % of their personal income was from freelancing:

  • Score 1 (never a contract): 30% of their income comes from freelancing
  • Score 2: 51%
  • Score 3: 45%
  • Score 4: 29%
  • Score 5 (always a contract): 15%

Molly: Do you ever consider usage in your negotiations? If you were making a map for commercial use for a real estate developer, that is a very different market than a map for an academic book. The price should be different even if the number of hours spent is the same.

Aly: Interesting question! I don’t think it specifically applies to much of the work that I do, but I definitely do think about if the client I’m making the map for will be using it for publication, or if they will be turning it around and selling it.

Daniel: I do a little. It’s mostly a gut calculation of how much money is worth to that client. Some people clearly have tight budgets, and others value money less and would probably be willing to pay more. I guess that’s more about the client than the usage, though.

Molly: Totally. Sometimes larger potential clients want to think of cartography as hourly work, whereas we’re more like illustrators on some projects. Making beautiful, informative works that help them reach their market. I feel like considering usage is one way to place more value on our work.

Daniel: Since I mentioned the percent of income freelancing question above, maybe now would be a good time to look at that.

I’m surprised there aren’t more people who are in the 90%+ bracket. I guess, looking around, I thought there were more freelancers who did this for most of their living.

A question that I’ve seen several people ask is: “Do part-timers charge less than full-timers?”

There’s not really a correlation here to be seen. Last survey there wasn’t one either, I don’t believe. When I see people asking that question, I get the sense that they feel that part-timers are underbidding full-timers, but I don’t think we see any evidence of that in either version of the survey.

Molly: When I freelanced on top of a full-time job, I leaned on so many assets of that full-time job—health insurance, namely, but also software and hardware. (Sorry!) It’s kind of crazy how much overhead is involved. Something to consider carefully when you go out on your own.

Aly: Absolutely, Molly, I assume many of us part-timers don’t have the flexibility to leave our full-time jobs because we so heavily depend on the benefits, primarily health insurance (for us Americans).

Daniel: One interesting thing I found, and for which I can’t figure out an answer: I looked at some differences by gender across some of the questions we’ve been discussing. Women and men were similar in their rate of using contracts, whether they were hourly or flat rate, etc. But, there was a difference in the percent of income question. Men received less of their income (13% mean, 28% median) from freelancing than women (48% mean and median).

Aly: Honestly, I wouldn’t be surprised if that had to do with inequality in (full-time job) salaries that meant women had to make up for more in their part-time work.

Molly: Interesting, Aly, about full-time salaries. I also suspect that women strike out on their own if they feel like opportunities for advancement aren’t equally open to them at their workplace. I would loooooovvve to hear from women about their interpretation of these numbers.

Daniel: Or even, if a man and a woman earned the same amount from a freelance project, it would represent a higher percentage of the woman’s total income if she were earning less at her day job than the man.

Aly: Exaaaactly.

Daniel: Which brings our attention to the table you have there about the pay gap, which unfortunately persists in the freelance mapping realm, too. And it appears to have gotten a little bigger than 2018, though the samples sizes are small enough that maybe that’s just noise. Definitely not getting smaller, though.

I took heart in one anecdote from the survey, though: of the six women who said they saw the 2018 survey results, five gave themselves raises (and one didn’t answer either way about a raise). The pay gap thrives on secrecy, and hopefully this survey will continue to bring some light.

Also, as a side note: 56% of respondents didn’t see the previous survey. So, audience: share with your freelance friends!

Molly: Data for the win!

Daniel: In 2018, we found, not surprisingly, that people who charge more are also more likely to say they are paid fairly. And at that time, women were generally less likely to feel they are paid fairly than men. Which made sense given that they were also earning less than them. This time, the pay gap persists, but it looks like overall satisfaction is up.

On a scale of 1–5, with 1 being “Never paid fairly for my time” and 5 being “Always,” women on average answered 4.1, and men 3.9. This is much closer than 2018, where women answered 3.1 and men 3.6. So, while the gap in pay hasn’t closed, the gap in satisfaction-from-pay seems to be smaller, to the point where it could just be noise in the data. And this whole batch of respondents for 2020 is happier with their pay.

Aly: Can we correlate any of that data with the people who gave themselves raises?

Daniel: For people who haven’t given themselves a raise in the last year, their average satisfaction score was 3.9. For those that had, it was 4.1. So, close, again. I wonder if people who see the results of the survey will change their opinion on whether they’re being paid fairly?

Aly: Here’s the median hourly rates broken down into different types of mapping:

Molly: Sighing, but not surprised, to see hand-drawn maps at the bottom of the barrel. I don’t know that there’s much to say about it though. I’m trying to re-frame it for potential customers: it’s a boutique service, not a hobby. (I don’t actually say “boutique service” out loud! It’s just something I repeat to myself when talking myself through an estimate.)

Daniel: That was very surprising to me. I feel like hand-drawn maps feel more elegant and bespoke. I was expecting them to command the highest price. I make custom maps digitally, but something hand drawn feels more obviously custom.

Aly: I agree, interactive maps require a separate “advanced” skill set so I wasn’t surprised to see them come in at a higher hourly rate, but I also consider hand-drawn to be a very specialized skill (I couldn’t do it!).

Daniel: Aly, I agree there’s a mindset that interactive mapping is “advanced,” certainly in the minds of clients, too, who are willing to pay more for something with cool animation, interactivity, etc. But I don’t want to sell short static mappers (which, I suppose, the three of us here are). I think we deal with different circumstances than interactive mappers, with different skillsets, but probably on average we spend the same amount of time acquiring those skills and thinking through how to apply them. I don’t have to figure out how to grapple with d3.js, but an interactive mapper doesn’t have to grapple with the challenges of getting something to look right on an offset press.

Aly: Very good point, Daniel!! I wonder how hand-drawn maps would fare against freelance illustrators, maybe that’s more comparable?

Molly: Great question. I do find myself relying more on the advice of illustrators when pricing and negotiating. There’s a lot more info out there re: illustration than cartography. Part of why you all created this survey!

Here are some illustration-oriented resources that I go back to again and again for advice on pricing, negotiating, etc.

  • Graphic Artists Guild — Guild membership gets you access to free webinars that can be helpful in running your business, and more. (This one by Emily Ruth Cohen on Advanced Pricing Strategies was particularly helpful to me.) Their regularly updated Handbook: Pricing & Ethical Guidelines includes charts of comparative fees for map design and illustration, as well as lots of information on contracts, negotiation, general practices, etc.
  • Whenever I need a pep talk on pricing, I revisit lettering artist and illustrator Jessica Hische’s The Dark Art of Pricing. It’s written with illustrators in mind but is so smart, borne from experience, and can help cartographers as well!
  • Map illustrator John Roman walks you through the business considerations of a project, from schedule through usage and credit.
  • Illustrator Anita Kunz gives great advice, especially for those newly in business. Her emphasis on keeping overhead low but also accounting for it in pricing makes for a great chat. (6 minutes)

Daniel: The cost of living was a new topic this time around, to get at a common question I’ve seen freelancers ask. If someone lives in an expensive area, I’ve seen them wonder if they could be underbid by someone who lives in an area where the cost of living is lower. So we asked people where they lived and then I looked up a bunch of those cost of living calculators online that people use to compare salaries between cities. I averaged several of them together to get a cost of living index.

There is no particular correlation here that I could see. There’s a slight incline to that trendline, but the R-squared is pretty low. So it looks like the two most common reasons I’ve seen cited for why a colleague might be able to underbid someone (living in a less expensive city, working only part-time) aren’t borne out by the data. Everyone’s rates are all over the place, without much regard to where we live, our education level, or other factors you might think have a role.

Molly: We should survey who is receiving coaching or mentoring. I swear that having a business therapist would really help me—but I can’t afford it!

Aly: There are business therapists?? That sounds amazing.

Daniel: I’ve never heard of those!

Molly: I don’t know but I want one, right?! So much of this is about learning to value one’s work and speak up for oneself.

Daniel: Sounds like many of us could use some help from one. I have been fortunate enough to have some colleagues to bounce ideas off of (like project rates), and who encourage me to charge more.

Aly: Which is the lovely thing about our cartography community, we have people who are always willing to share their expertise (not to mention sending projects your way!). I’m still always impressed how much of my work is word of mouth through the cartography community.

Daniel: A fair bit of mine has been, too! When people ask me about freelancing advice, I don’t have a lot of good specifics to share, but I do tell them to get connected into the community, because they’ll learn stuff to make better maps, and they may also get leads on work.

Molly: Oh! That might be a good question for the next survey: what percentage of your work comes from where? (e.g., word of mouth through friends/past clients/other cartographer; via website; via NACIS list, etc.).

Aly: Definitely! New freelance cartographers always ask me how I get clients and I don’t have good advice outside of getting connected to our community. If I did this full-time I’d have to do a lot more work putting myself out there.

Molly: Did we get an overall result for who has given themselves raises since 2018?

Daniel: Of the 54 people who answered the question, 26 (48%) said they’d given themselves a raise. That’s actually more than I might have expected.

Most people’s answers of how they determined when to raise their rates were either based on external factors (based on what they make at their salaried job, or what their regular client negotiates), or were more ad-hoc, as best summed up by one respondent:

“Whenever I think I can get away with it.”

It’s something I don’t do often enough, but since the 2018 survey, I’ve been thinking about. I raised them after that survey, and am keeping it in mind more in the future.
Thanks, both of you, for taking the time to chat about all this stuff! I feel like it makes it more interesting than just dropping a bunch of numbers and charts on people. Gives some context and interesting side points.

Aly: Absolutely, thanks team! This was great!

Molly: Thank you for the chance to join you! I loved it. Feel free to edit me liberally.

(Note: all of us were edited liberally)