Atlas in A Day

Atlas in a Day is a mapping challenge run by Guerrilla Cartography

Cartographers around the world wait on tenterhooks for a theme to be announced, and then have just under 24 hours to design, create, and submit any kind of map on that theme. The maps are collated into an atlas and published at the end of the challenge.

I was asked by Andrew Douglas-Clifford if I wanted to collaborate on an entry. I gladly accepted the invite.

Atlas in a Day took place over the weekend of 15th – 17th May 2020, and the theme was…


CommunityCredit: Guerrilla Cartography


After an hour or two of brainstorming ideas, we settled on mapping the global locations of all Sister Cities of New Zealand. According to their website

“Sister Cities aims to foster cross-border communications

– a mutual exchange of ideas, people, and materials in a range of cultural, educational, youth, sports, municipal, professional, and technical projects.”

We felt that fitted the theme quite well.


Our list of Sister Cities came from the website itself, where there is a separate page for each foreign country with tables listing the city pairings. In total there were over 150 pairs.

To create a geographical dataset of these cities, I used Google Maps to search for each city (copying and pasting in one name at a time) and then eyeballed the location in ArcGIS Pro and created a point including attributes for Name, Country, and Region. Was this a tedious process? Yes. Was it the quickest process? Also probably yes.

I initially thought about writing a bit of a script that would could call an API and geocode the names. I’m very glad I didn’t go down this particular rabbithole, as it probably would’ve taken longer to write something that would reliably geocode all of the city names.

Google Maps

Some of the names (particularly Japanese cities) didn’t return any search results.

After searching these missing names in Wikipedia, I was informed that they *were* cities, but had been amalgamated with others through urban expansion and were no longer known by those names.



As this map was about communities related to New Zealand, we wanted an NZ-centric map.

We thought about having everything on a single projection, using some kind of stereographic projection, but decided against this as it would introduce huge distortion to countries on the other side of the world, and there are multiple Sister Cities in Europe.

Instead Andrew suggested having several region maps dotted around the central NZ map.  He illustrated his idea with a sketch.

A very rough first sketch of the proposed layout.  Credit: Andrew Douglas-Clifford

Once the city locations had been digitised, it was simple to identify clusters and split them into Europe, Asia, Australia, Pacific, and North America.

For each region, I created a new map with the “World from Space” orthographic projection, centred roughly in the middle of that region’s cluster. Each region’s points fitted nicely within a half-diameter circle*, so the resulting maps could all be the same size and scale.

*A little cartographic license was employed when we shifted Hawaii about 3,000km east to fit in the map with the rest of North America


It became very apparent that we wouldn’t be able to show every city-to-city connection.  In some places they were far too closely packed that it would just become (even more of) a mess.  We considered only showing the connections from the major NZ centres, but decided this would do a disservice to the numerous small towns in the regions.  Instead, Andrew suggested we create the links from NZ locations to the world regions, but not to individual global cities, and we thought this was a decent compromise.

Sorry Japan, there’s just too many of you!

Each NZ region gets its own colour

To provide a bit of distinction between the lines, we decided to colour-code them according to which NZ region they connected to.

Andrew also tallied up the connections for each world region in order to label the nodes.

Working in Adobe Illustrator, we each took a world region and added the swooping lines.  We tried to make them as smooth and flowing as possible, and avoid overlaps from the same world region if we could help it.

Several of the lines became trunk-and-branch, where lines from several NZ cities to the same world region would naturally blend together.  There were obviously going to be overlapping lines where different world region connections crossed.

We could easily have spent another 24 hours just tidying up the lines, but we only had about half an hour left and were happy enough with what we had.

Trunk-and-branch lines


Most of the hard work was now complete.

Andrew created a title graphic for the map, with swooping lines connecting the tittle on each ‘i’.

We added a final bit of text explaining what Sister Cities were all about, and the map was complete.

We submitted it with two minutes to spare.

We were required to submit our map as a rasterised PDF.  Neither of us could remember how to do this.

For future reference when I inevitably forget how to do this again, our frantically Googled solution was to export to vector PDF from Illustrator, open that in Photoshop, and re-export as a Photoshop PDF.  There’s probably an easier method, but this did the trick.

Our completed map.  The full Atlas in a Day: Community can be viewed online here.


I had mixed feelings when I looked back at our map after we submitted it.  My first thought (which was a recurring thought for both of us during the whole 24 hours) was that it was a bit of a mess.  There really are a lot of lines!

But on further reflection, I’m actually very pleased with our map, and what we achieved.  In under 24 hours (a large proportion of which was spent sleeping) Andrew and I successfully created a map on a theme that initially had us flummoxed.

I’m quite proud of that.  Let’s do it again!

Out thoughts throughout

Comments are closed.

hide comments

Your cart


No products in the cart.


This is a unique website which will require a more modern browser to work!

Please upgrade today!