——————————————————————————–
Note 8/3/2016: Not long after writing this post I came to my senses and revised the coloring book. “Ha ha,” you say. “After all the defensiveness in this post you actually revised the book!” And you are right to laugh, but let me explain.

As is typical in cartography and perhaps all subjective media, the criticisms thrown at the book were highlighting a deficiency but not quite hitting on the exact problems. In this case, after an immense amount of thought and hang-wringing, I realized that the real problem was with the linework. And so I fixed that linework to the best of my ability and released a revised book interior in July 2016. The details on that revision are here.
———————————————————————————

Just me, #CityMaps, one colored pencil, some gin & tonic and another 30 min left in this flight. pic.twitter.com/RqnvvJM5PX

— Sarah Wray (@SunWray) May 13, 2016

A few constructive criticisms have surfaced regarding the coloring book City Maps: A coloring book for adults that I released less than two months ago. I’m going to address those criticisms in this post.

In the less than two months of its existence, it has done phenomenally well as far as I’m concerned, with more than 2,000 copies sold and write-ups in The Atlantic’s City Lab, GIS Lounge, GIS User, Curbed and more to come in the months ahead. Furthermore, it went a bit viral over on Facebook following the City Lab article for at least a week to the tune of 14,000 shares (say what?!).

For this book to have sold 2,000 copies around the world in less than 2 months is amazing and is what I’m focusing on in terms of whether I consider it a success or not (I unequivocally do). Check this awesomeness out:

And at one point it was near Harry Potter in sales! Ok, for just a few days, but heck yeah for coloring books!

However despite very strong sales, the book just recently got some very intriguing criticism from a few people that I’d like to address.

We’ll disregard a couple of the critical reviews that were just outright scammy and focus on the few that had specific issues with the book.

As far as the honest comments were concerned, one said the maps appeared to be copied and pasted in the span of 30 minutes, one indicated that they thought the maps weren’t accurate, and another discussed some feature size, labeling, and dangles issues.

Map Making Procedure

Starting with the comment on procedure, which purported to be from an urban planner. I’m not really sure how a person would copy and paste map data as this person thinks, but I can assure you if that were possible then a book like this would have been created long ago. I’d like to take a moment to describe the exact procedure for creating the map pages, since I think it is instructive for those who read this blog for tips on map-making as well as addresses this criticism specifically.

The book was created in a highly focused manner over an extremely intense span of about 80 hours plus revision time after that. During those 80 hours, OpenStreetMap extracts for all the major metropolitan areas of the world were downloaded. My familiarity with OpenStreetMap data helped in this regard because I knew how to obtain it, in what format (I used osm2pgsql data for this) I wanted it, and what the fields and tags mean as well as how to query for the right mix of data such as primary roads over residential or querying out urban area polygons so that they didn’t appear, and so on.

All this data was imported into separate QGIS projects where suitable locations were chosen with the following criteria: (1) good density of information (OSM data is not always complete in some locations) (2) a place that made sense to color either from a famous-landmark perspective OR from an interesting-shapes perspective. These locations were researched to determine a place name (the beginning of the book caveats that place names are subjective nevertheless as locals may have different names, but these names are adequate for the layperson to look up the location on online mapping platforms if they are interested in what a particular shape represents, for example). They were also researched to ensure that they were adequately representing the area and depicting something interesting and, hopefully, non controversial.

A local projection for each location was researched and applied to each project. The data were massaged so that a good mixture of spaces to color and lines representing real-world locations came though. This was not always easy to get exactly right and I think that in some places it may not be what people would expect and could be better. Each image was exported at 600 dpi for inclusion in the final book.

The final portion of the procedure was book layout. I was aided in the fact that I’ve published before and know the ins and outs of layout. So, because this book was very simple layout-wise, I was able to do this in the easiest manner possible, using a pre-formatted template in Word that was designed specifically for my two destributors: Ingram (everyone but Amazon’s supplier) and CreateSpace (Amazon’s supplier). Cover files, which include front, spine, and back, I created in Inkscape to the exact specifications for each printer (they are different for each) and created for the width appropriate for the paper thickness and the number of pages so that the spine fits almost exactly. Due to printer error-margins the spine has a bit of error-space as well.

I already have accounts with my two major suppliers – Ingram requires quite a few signed documents that I was thankful to have already taken care of years ago – so the process for handing over the interior and cover files was straightforward for me. A new person to this process has a much harder time so this definitely saved me a lot of time. After that it is a matter of waiting for proofs and revising those proofs until I was happy. For example, initially the width  of some of the line features was not adequate on the presses. I increased the weights for some, decreased the weights for others, and just in general evened things out if there were large discrepancies between feature width on the same page for many of the pages. Though it must be noted that I did leave some variations in line-widths to comport with what you see in existing adult coloring books where there are some thick and some thin lines to create a visual variety.

With that I think I’ve adequately addressed procedure and perhaps shown that it would have been absolutely impossible to “copy and paste” a book like this together. Next up is the accuracy issue.

Accuracy

A critic said that the maps lack accuracy. Let me speak to that as best I can as I don’t have any more information regarding what exactly they found lacking in this regard.

The copyright page lays out, perhaps too succinctly, the necessary disclaimers regarding the mapped data. The copyright page, in retrospect, may not have been the best place for this information as it may easily be overlooked. I squeezed all the text information onto that page in order to reduce page count so that I could maximize the coloring pages while keeping the cost quite low – at $9.99 US.

The copyright page points out the following:

Information printed in the front matter.

That information states very clearly that the map data is from OpenStreetMap, indeed that what we are talking about is “data” and is as accurate as the chosen dataset. In my professional opinion, the OpenStreetMap data is reasonably accurate in the locations that were chosen for a coloring book. The aim for the book was to include at least one map from all of the top 10 largest metropolitan areas of the world, though in the end I did have to leave out Osaka due to a lack of a proper density of data for coloring. All the other metropolitan areas of the world are included, plus a smattering of other interesting locations such as Boston, Vancouver, and Venice. Ensuring that all the most populous metro areas were included was a way of making sure that I covered the globe, so to speak. There are other “great cities” books which focus on only those cities that Western  cultures are familiar with traveling to and from and I wanted to be more representative of the real world.

That’s why, if you are from the U.S., you might be surprised to see that a coloring page for a portion of Manila is included, for example, as this might be a place that isn’t high on your “to visit” or “have visited” or “have studied” list of cities. But it is one of the largest metro areas in the world. Also, I had at least one book buyer who was from Manila and who was happy with this inclusion.

In terms of accuracy, I first wondered if perhaps the critic might be too used to seeing Google Maps, which uses the Mercator projection. The maps included in the City Maps coloring book are each projected to a local coordinate system that is appropriate for the location, not Mercator, which is notorious for its inaccuracies with regard to shape and area and is used in webmapping systems primarily because it became the default projection for them after the first webmapping platforms employed it as an easier mathematical solution to their map tiling needs.

Let’s take a look at this in detail. The very first map in the book,The Palace Museum in Beijing, China (also known as The Forbidden City), is shown here in thumbnail form:

Thumbnail version of the first coloring page. Beijing – The Palace Museum

A close-up of the southeast corner of this map is shown below:

This map is in the projection Beijing 1954 / Gauss-Kruger zone 20, also known as EPSG: 21420. The Mercator equivalent of the same location on Google Maps appears like this:

Mercator Aerial

In this particular location and scale I don’t see too much difference between the projections and I see that the mapped data from OSM appears to be fairly accurate, though of course not every single feature is mapped.

In the coloring page close-up we see a few issues with overlapping lines at the southern portion of the page. This is definitely an issue that may be cleaned up in a future edition. It’s an artifact of OSM data wherein many people may map different cultural tags for the same location. One person may have mapped just the general location of The Palace Museum with an outline, for example, while another OSM contributor may have mapped the individual stalls. Many many of these artifacts were cleaned during the production of the book via teasing out the individual “tags” (in OSM parlance a tag is how you describe what the mapped location is) such that things like urban areas or shipping lanes, for example, weren’t included. They tend to confuse the map reader, who doesn’t expect to see lines in the water even though they do signify real things. I do think that further cleaning of this sort is warranted and I hope to ameliorate this in the future. Does it preclude your use of the book for coloring? That’s for you to decide.

Now let’s take a look at Manila since I’ve mentioned it already. That particular location was difficult because the OSM data for it wasn’t as detailed as in other locales. So this map is at a smaller scale and includes line features only where “water” = ‘river’ (not road lines) and all polygon features. The interstices of the polygons reveal the road network at this scale. A thumbnail version of the page is shown here:

Manila – Port Area

The southeast corner of this map looks like this:

Manila Port Area southeast corner

This map is in the PRS92 / Philippines zone 3 projection, also known as EPSG: 3123.

And here’s the aerial of the area (Mercator projection, from Google Maps):

Manila Port Area, Aerial

Here’s the Google Map of the same location:

Manila Port Area Google Maps

I think that the projection differences at these scales may not be enough to cause someone to say that the maps are “inaccurate” so perhaps my thesis on this being the issue is incorrect.

At any rate there are certainly some differences between OSM data and other mapped data, and that may be where we come into conflict with what people might deem to be “accurate” vs. “non.” When we do compare the maps shown above, though, we see a close-enough association, in my mind, for the colorer.

A NOTE ABOUT AUDIENCE I’m bolding the word colorer because this is where I want to talk a little bit about the audience for the book: the colorer. I’m forever talking about “know your audience” when it comes to creating a cartographic product. We must ask ourselves what our audience is expecting, how they need it to be presented for maximum informational absorption, and what we can bring to the map that adds to the audience’s experience and knowledge-base in a positive and efficient manner.

After hounding this in to countless people–most recently up in Manitoba where I gave several talks/workshops to the municipal government and planning association–I’m ashamed to say I may have misjudged the audience for this book given some of the feedback I have received! I believe that the primary audience is the mindfulness adult colorer and I still believe that. What I failed to grasp was just how many people would become a secondary audience for this book: those who are map nerds/cartographic enthusiasts who will  be sticklers for accuracy at the expense of coloring quality. For this secondary audience I should have made very clear, in a special section at the beginning of the book, exactly how the book was made, what the potential pitfalls are, and my reasoning behind not including labels, choosing the locations, constraints, and so on.

Small spaces, labels, and dangles

One criticism was actually with regard to the primary audience: those who want clear spaces to color. This critique focused on the fact that in some of the maps there are some spaces that are too small to color. In particular I point to an example from the Bidhannagar map:

Small locations on the Bidhannagar page, actual size, Asia South Lambert Conformal Conic, EPSG: 102030.

Bidhannagar Google Maps, Mercator projection

When you compare the two maps, you can see in the coloring book page that there are fewer roads. This is because I only included polygons on this map. In this case I felt the polygons adequately represented major roads in their interstices, just as with the Manila coloring page, and also that including roads in this location would have caused a huge density issue (way too many small spaces to color). A few of the smallest polygons shown in that snippet at the above-right, are indeed too small to color. A closer look at the data reveals that these particular polygons are all tagged as water. A filter to get rid of polygons smaller than a particular size would have been in order here.

This was a compromise on my part having to do with the amount of time I had available to me and it may have been a poor decision. I do know that at least one person has shown me her colored version of this page and she simply colored right over these smaller boxes. The lines show through the color and seem to still look okay to me. I know from reading hundreds of coloring book criticisms that there are colorers who don’t mind small spaces, some who abhor them, and some who criticize if the spaces are too large (i.e., more like a child’s coloring book). It seems very difficult to get it just right on this score and I hoped that the compromises I made here were good enough. That said, it may be something to take a closer look at in a second edition if I do one.

Labeling of the maps is something that has come up a few times from a few cartographers I know. The maps in this coloring book are not labeled and this was a purposeful decision, again, with respect to the audience primarily being concerned with coloring. I didn’t include labels since those really get in the way of the coloring experience. I’m inclined to disregard this particular notion because for this primary audience of the adult colorer, I think that labeling would not be good.

Line dangles are definitely an issue in some of the maps. They were very difficult to get around considering the scale of the maps, the data, and the fact that many of the maps feature extra-wide road casing in order to make the roads colorable. These road casings do stick out into water locations on some pages. In another edition these definitely need to be cleaned up manually in Inkscape. I’ve wavered over whether to pull the book from the market to fix this particular issue, but I am convinced that the ROI on this is not enough to make this worthwhile. If there comes a time when I want to completely update, add new maps, and so on, then I will address this. In the meantime, I think that for the price, quantity of maps to color, and the interesting places to explore (I know many who have looked up the maps to discover more about a place), this book is more than worth the price. I’ve thoroughly enjoyed coloring most of the pages myself despite a few line dangles on some of the pages. I could fix them but considering my already too-booked schedule of consulting and speaking I just didn’t find that it would be worth it for the small increase in satisfaction that it would give to the sticklers. I’m thinking I may slightly edit the book’s description on the Amazon page to ensure that nobody incorrectly buys it thinking that it is a detailed road atlas or anything remotely similar.

City Maps constitutes a mash-up of two popular genres: adult coloring and cartography. This concept had never really been put forth in book format, in our recent past, before*. As such I believe these few critiques stem from the fact that it is a unique and interesting type of book and this kind of unique product will always bring with it a healthly amount of skepticism from those who may have expected something a little different. The hundreds of privately and publicly posted expressions of enthusiasm, support, and positive critiques has helped greatly to outweigh these issues addressed here but no good cartographer should create a product without a thorough look-back and lessons learned for the future as I’ve done here.**

*Apparently someone created a city maps coloring book 400 years ago?

**I propose that cartographers in production work incorporate some sort of “code review” and “retrospective” process (terms borrowed from software engineering).

hey @PetersonGIS my (almost) five year old is also loving your map coloring book. have to get my own I guess… pic.twitter.com/pLz6RHRtxH

— Doug (@dropstones) May 21, 2016

.@ArcGISOnline has its place, but it shouldn’t be oversold. Or people will get frustrated. Like me. I’m coloring. pic.twitter.com/cmgAYaLlNz

— David Puckett (@BrightRain) May 5, 2016

***********

Wherein I address a few concerns about “City Maps” & provide a detailed discussion on how it was made! https://t.co/0DwUfSWUzH

— Gretchen Peterson (@PetersonGIS) May 23, 2016

Pretty sure that was my longest blog post ever. In 6 yrs of blogging. And it was maybe the 1st time ever that it was more about me than u?

— Gretchen Peterson (@PetersonGIS) May 23, 2016

@PetersonGIS I enjoyed reading how you made the maps for the coloring book, and was SMH at the people who gave you grief.

— Barrett Clark (@barrettclark) May 23, 2016

@PetersonGIS It was an interesting read, especially for someone like me in the secondary audience. 😉

— Robert Nordan (@robpvn) May 23, 2016

@PetersonGIS osm2pgsql? Didn’t you used to be a die hard imposm user?

— Craig Williams (@williamscraigm) May 23, 2016

@williamscraigm Good catch. osm2pgsql keeps it in points, lines, polys–just 3 separate datasets–but can be too big for world-wide that way

— Gretchen Peterson (@PetersonGIS) May 23, 2016

@williamscraigm Since this was just city extracts I didn’t mind that this time around. imposm splits into however many tables you specify.

— Gretchen Peterson (@PetersonGIS) May 23, 2016

Apparently there is a vocal market for spatially accurate map coloring books, suitable for litigating boundary disputes, well siting, etc.

— Brian Timoney (@briantimoney) May 23, 2016

@PetersonGIS Oh dear how ridiculous you had to write such a response!

— Nathanael Coyne (@NathanaelB) May 23, 2016

I cannot believe that @PetersonGIS has to defend the spatial accuracy of her cool City Maps coloring book pic.twitter.com/27Unz2REnT

— Silas Toms (@silasmerlin) May 25, 2016

The spirit animal coloring book needs to explain as to why there is no snuffaluffagus! @silasmerlin @PetersonGIS https://t.co/uzwdbX4M9e

— Zac Andereck (@ZacAndereck) May 25, 2016