How many people live in Tokyo?

Back in August 2014 when I was preparing some material for teaching students how to query data in a GIS, I devised a very simple example where they had to select all the countries in the world with a lower population than the Tokyo Metropolitan Area (or Greater Tokyo Area, as it's sometimes called). I did this just as an example but since I found the results quite interesting, I quickly turned it into a map and posted it on Twitter, complete with 'Toyko' typo in the subheading. It was really just a quick example of how to query data in a GIS but it also highlighted the massive population of the Tokyo metropolitan area. You can find the full size, original image here. I was prompted to write this blog because the map was reposted on the Canadian Twitter feed @AsapSCIENCE a few days ago and since then my inbox has been a bit busy.

Now, back to the original question of how many people live in 'Tokyo'. Well, when I say 'Tokyo' in the map, I'm referring to the wider metropolitan area, which in 2014 the United Nations World Urbanization Prospects said has a total population of 37,833,000, far beyond the next largest urban agglomeration, Delhi, at 24,953,000. By way of comparison, using the urban agglomeration population (rather than a city's administrative boundaries), Toronto had 5.9 million, London had 10.2 million and Beijing had 19.5 million. Clearly, these definitions include other urban places that are not Tokyo (e.g. Yokohama) but they are recognized as being part of a fairly coherent metropolitan area. 

So, in the map, I'm using a figure of 'approx 36 million' as I say but in reality the UN figure is a bit higher. If we want to narrow the definition down to just the inner urban area then obviously the figure reduces significantly. I'm not usually one to cite Wikipedia, but in this case it's a good place to go to learn about the various definitions of the Greater Tokyo Area/Kantō region. If you don't want to click, here's a summary of some Tokyo populations...

• Former city area (23 'special wards') - 8.95 million people
• Tokyo Metropolis - 13.05 million
• Tokyo Metropolitan Employment Area - 31.70 million
• National Capital Region - 43.47 million

If we take the metropolitan definition used by the United Nations, then Tokyo does indeed have more people than Canada, at 35.5 million (2014 Statistics Canada estimate) and far more than Australia at 23.7 million (see their population clock). The Tokyo metropolitan population is roughly the same as the whole of California, which currently stands at about 38 million.

Link to chart

This was all just a little bit of map trivia and whilst it seems to have annoyed some people who live in a red country, the point was just to demonstrate the simple analytical power of GIS in addition to the size of Tokyo (to make it more interesting). The data I used are from Natural Earth if you want to have a look yourself and the software I used is a free GIS called QGIS, which is really good. Some other random facts about what happened to my original tweet...

• It was tweeted by Guardian Cities and then retweeted 309 times and favourited 107 times.
• It was tweeted by AsapSCIENCE on 22 January 2015 and since then retweeted 96 times and favourited 110 times.
• Akihabara News wrote an entire article about the map, and even spelled my name correctly - that hardly ever happens.
• Tokyo Desu also did a piece on it - and added another map too.
• What does it all look like on the ground? This photo (below) from ATBREAK gives you some idea.

Tokyo metropolitan area

World city rankings by search engine results

I've been thinking recently a lot about methods we use to rank, count and understand urban areas. This is a topic which the Globalisation and World Cities Group (GaWC) at Loughborough University in the UK have become world leaders in. I also wrote a paper about this some time ago and I'm still interested in it more generally but not so much from an academic research point of view these days. However, I was reminded about a paper which I referred to in the SGJ article mentioned above - a paper written by the aptly named Florian Urban. He looked at the representation of cities on the internet in his 2002 paper from the perspective of place marketing and representation. However, I wondered recently what outcome you would get if you simply entered city names into google (I did try others but google seems to return more results than all others). Not surprisingly, the results match those which are typically published and based on much more complex analyses...[read the caveats below too]

1. New York = 827m ("new york city" returns only about 102m)
2. London = 539m
3. Paris = 597m
4. Tokyo = 142m in English, plus 427m for 東京 for a total of 569m
5. Hong Kong = 246m in English, plus 290m for 香港 for a total of 536m
6. Los Angeles = 267m (quotes or not, doesn't seem to make much difference)
7. Singapore = 221m
8. Chicago = 337m
9. Seoul = 35.5m
10. Toronto = 157m

Caveats... New York could be the city or state, Seoul has some name issues that mean even if you use the simplified Chinese characters to seach the indexed pages result doesn't increase much, there is a massive English speaking bias here too (but then that reflects the real world to an extent). Also, the extent to which the web is US-centric comes out here. For comparison, Beijing/北京 returns (105+555 =) 660m. You'll also get different results day to day, with a general trend of increase.

Okay, so these results are not at all scientific, but they do quite closely match the kinds of rankings you see in the academic, popular and web press. What about cities in the UK? Well, we have Manchester at 139m ( swelled by football references?), Birmingham at 85m, Glasgow at 58m, Edinburgh at 53m, Leeds at 46m, Sheffield at 39m and Inverness at 14m. No doubt there's a lot of noise in these results but again it mirrors other ways of measuring.

In my work, I'm much more interested in the places far lower down the urban 'hierarchy' and not so much in rankings but it's interesting to see how this little exercise has turned out.