If you regularly visit this site, you know I’m a big fan of Tile Grid Maps. They are relatively easy to make (in Excel, at least), engage readers in a different way, and address some of the issues with traditional choropleth maps (though introduce their own issues).
Following up on the Tile Grid Waffle Chart Map I made of Africa a few weeks ago–and a tweet by a certain Ann Emery–I decided to sit down and try to make a Tile Grid Map of the entire world.
Has anyone made a #TileGridMap of the entire world yet, @jschwabish? (@visualisingdata, want to add this to your to-do list please? ????)
— Ann K. Emery (@AnnKEmery) September 22, 2017
I’ll admit, it was harder than I thought. For one, I’m no geography expert. Second, some countries have really weird shapes–Vietnam is super long. And third, getting the squares in just the right layout was really hard. If you dig into your area of the world, I’m 100% sure you’re going to find something to argue with me about.
In the end, I’m not even sure this is a useful visualization type. Take the case of Russia. In case you don’t know, Russia is a HUGE country. A quick Google search yields the following estimates of square miles for the five largest countries in the world:
| 1 | Russia | 6.6 million square miles |
| =2 | Canada | 3.8 million square miles |
| =2 | United States | 3.8 million square miles |
| 4 | China | 3.7 million square miles |
| 5 | Brazil | 3.3 million square miles |
In other words, Russia is nearly as big as Canada and the U.S. combined! Have you looked at a world map lately? Russia stretches from eastern Europe, past China, nearly to Alaska!
So is a map where all countries are equally-sized squares being fair to a country that’s so large and spans such a wide area of the planet? I’m not sure. On the one hand, it actually makes the point stronger–visualizing data for which the value for Russia is not meaningful, the WorldTGM minimizes the visual space Russia takes up on the map; instead, it allows emphasis to be placed on other countries for which the variable may be more meaningful. On the other hand, it dramatically changes the visual of the world and makes placing the tiles in reasonable proximity to each other virtually impossible.
Anyways, I’m providing my version of the WorldTGM in Excel format here. Because someone asked, I’m releasing this file for free under a CC-BY-SA license, which means you can use and adapt it, as long as you give attribution and share your version under the same licensing.
You can download the file and move the squares around to your heart’s content. If you come up with an alternative, please post it below–I’d love to see the different versions! And maybe Jane Pong can make another cool animated gif of the different approaches!
I’ve also made a version that pulls in real data, and am now working on the labels; I hope to have that out in a few weeks. In the meantime, enjoy this one!
UPDATE: Mustafa Saifee has converted the World TGM to a JSON file, which you can get from his Github page here.
Hi, big fan.
I’m curious about your process for this, as I’ve been working on one for counties of NC. I’ve tried both by hand with an overlay, and writing a short script based on centroids of the polygons in R. How do you actually DO this, when you’re doing it, in a way that doesn’t take an eternity?
Hi Mike,
Thanks for writing! I basically do it all manually, sad to say. I’m sure there’s an algorithmic way to create the map, but I basically go country by country, trying to lay it all out. I try to blitz through the first pass and just get all the countries down somewhere close to where I think they should be. I then most of my time moving them around. Sorry, no magic answer!
Thanks again,
Jon
Hi, Jon:
I like to just admit when I’m beat when doing tile maps. For instance, any tile map I do of the US, I just grab DC and put out in the ocean. This is a standard choropleth practice when using geographic shapes–I don’t see why I can’t use it for a tile map!
My choice in Europe is to pull out the micronations and put them in a block off the map for the same reasons, and also Russia for the exact opposite reason. I think the Carribean nations probably would work this way as well.
I think this strikes a good balance between geographic and non-geographic representation.
Haven’t had the chance to do any others besides Europe and the US states, and I’m a degenerate hexagon tile user, anyway, but I love this conversation!
One way to deal with big and/or populous countries would be to make a tile for each of their biggest subnational administrative subdivision: i.e. replace the US tile with 50 US state tiles and so on. One could also editorialize and use non-administrative regions or something like that if that led to too many tiles.
Hi Jon,
I made a world tile grid recently for my Information is Beautiful entry;
https://www.informationisbeautifulawards.com/showcase/2479-the-state-of-the-nations
“It was harder than I thought” would sum up my experience with this too. I started on the assumption that grabbing the country centroids, plotting them in Excel and then using ranking, sorting etc. would give me a way to shuffle them around in a near-automatic way. It didn’t. So, assuming I was just being a bit dopey, I asked the Quora community for advice. I had an advanced mathematician say, basically, it’s really hard. As you’ve pointed out, and I found out through repeated attempts, it’s about choices and compromises. How compact are you trying to make the map? How important is relative position between countries? And so on. You’ve already highlighted the hardest bit to solve; the massive difference in size between countries gives you over-populated or disconnected regions.
Eventually I found a grid size which got me 70% of the way to what I wanted; a mostly compact map which allowed some spacing to maintain a familar feel to traditional maps. That is, I chose not to compress the East Asia and Pacific region because that distorted the relative positions of the countries, whereas I was able to force the other regions to mostly adhere to compact grids. The rest I did manually by looking at a traditional map and adjusting my tile coordinates to close gaps or spread small countries out.
I’ve put the Excel file on my OneDrive should anyone want to use it. It includes a scatterplot of the map so you can play with the coordinates and see the change.
Thanks for your site and the podcasts; they’re truly inspirational!
Attached image:
Very cool! Can you provide the link to your OneDrive so I can take a closer look at your layout?
Thanks,
Jon
Thanks!
https://1drv.ms/f/s!AiM4koV3zAHCgehRhKXq0TGolw8olg
A similar map was created by Catherine Reeves and published in Globehead! A Journal of Extreme Cartography in 1994. It eschewed the strict grid in order to preserve some local geographic relationships.
I’m pretty sure the original journal was paper-only, but it’s referenced here: https://bigthink.com/strange-maps/201-the-reeves-equinational-projection/
Very useful map for my work, thank you! Just wanted to point out Georgia is there twice, and I added Andorra, San Marino, Liechtenstein and Monaco for completeness.