Mapped Web
Physical distances are not the same as psychological distances. Physical distances are easy enough to measure, but how do we go about measuring psychological distances? The Mapped Web does this by taking the chance that given a page contains the name of one country it will also contain the name of another country as a measure for psychological distance. The resulting images show us how close countries are to each other in psychological terms.
This procedure can be extended to create maps for specific areas by including extra search terms. Look at all the pages on the Internet with the word War on them, which countries correlate? Probably countries that have something to do with each other in wars. The images below picture some of the relations between countries in this way.
A 200x200 matrix with the relative frequencies of country name combinations represents actually a 200 dimensional space with the countries as dots in between them. The trick is to reduce this amount of dimensions to two so we can actually plot something. This is usually done by a technique called Principal Component Analysis, where two new axis are constructed that represent the lot in the best possible way. The program here uses a different technology, the sammon algorithm. Here we just start with a random initial configuration. Over a number of iterations, we move the countries in the directions where they want to go, i.e. they are attracted to countries which they are closer to according to the matrix than according to the configuration and are repulsed by countries are too close to.
The results are the maps below. You can download two scripts, generateMatrix, which takes one parameter, the keyword, and generates a text file with the distances matrix for that keyword and stressOptimize, which generates the map. The latter uses the brilliant VPython visualization. The nice thing is that it is animated during the iteration, so you see what goes on. This project was largely done together with Ernst Wit, a fellow Savage Minder, who also came up with the initial idea.
The World Map according to the Web
The upper part of the map is dominated by the Anglosaxion countries; To the right Europe sits, with Germany and France in ever closer union. The lower part of the map is Asia, with Japan the most central because westernized nation.
Economics relations in the world
In the middle, the powerhouses of the world, the US, Japan and Germany, with China and its two satelites Hong-Kong and Taiwan waiting in the wings. What about the cluster India, Russia and Israel? I don't really know, but they do correlate. May because they are relatively poor countries that do well on technology?
Wars of the world according to the web
The war map is spread out and tense; so many wars, so little space on this bitmap. The second world war axis are visible in the right corner, as are the axis of evil in the lower right corner. Russia, China and Japan form an uneasy triangle of historic conflict. The US is in the middle of everything.
Immigration streams
Ireland and Germandy are in the middle, as large contributors to American immigration, but destinations themselves nowdays. The big immigration countries keep their distances from the others and warp the map a bit; immigration isn't that much an issue there as it is for the exporters of people in Asia, who are much closer to each other.
preview
downloadable files
no title
by Ryan
Seems to be pretty accurate, considering the way it must collect the data.
fsd
by sfd
sfd
Brilliant
by Z.
Quite an excellent idea, and works out nice in practice, too. Really enjoying the whole Projects page.
i am just studying geography ...
by Steve
... and all this stuff looks very interesting. thumbs up
Sweet!
by Paul B Hartzog
Douwe,
this is sweet, can you send me the whole thing?
-paul
check the record, check the record, check the guy's track record
by chaizzilla
> The US is in the middle of everything.
tru dat.
have you seen the "purple haze" maps going around lately?
What software do you use?
by Martin
This is a GREAT idea to visually represent flux of goods and people. Well done.
I need to create some mapsphere (of the global view of the local residents. Antrhopological research blabla... ) And this Idea is Great one.
Any chance you could send me info?
Good, but it needs something...
by
... Like African, Central and South American countries. The countries are in the generateMatrix.py code but I'm no coder so I'm kind of amazed the results leave out about half the world.
Very Intersting
by Vits
It's a great idea about how to present the correlation between to contries, by the way, it plots what the Algorithm sees, does it see the truth??
Nice work man,
Keep on it, you'll be rich some day...
Vits
Clever
by Steve
Very very ingenious and brilliant! This is one of the coolest ideas I've ever not thought up! Really clever - keep up the good work!
Spatially Mine!!
by Chris Adam
This kinda mapping takes me back to my 'spatially mine' Geography courses and our Theme Song of the early seventies!!
Blair's World
by Chris "Biff"
After all of our investigative efforts, where is Blair's World in these scenarios???
The psychological distance between EARTH and matter was defined at the Rhode Island Convention. Does anyone remember?
Language?
by Arthur J. Schotgerrits
Hi, I have one question..
Did you take all the different languages of all sites around the world in your equation or just English phrases?
good one!
by anne
hey its reall a great idea to corelate the countries on a number of basis, i found it very interesting but i have no idea of vpython
Sammon algorithm
by rdm
Sammon algorithm--is it the same as multidimensional scaling?
These plots are really interesting. Not sure if all of them make sense.
As an ecologist, I use similar plots to show how species co-occur in the environment. It's cool to see it used in this context!
Canada and Afghanistan
by Jason
It would be interesting to track the relationships between countries over time and present them as an animation.
Canada is HEAVILY involved in the war effort in Afghanistan and I was surprised that it was positioned closer to Iraq than Afghanistan (since we have no forces in Iraq now). I imagine if you tracked the changing relationship you'd see Canada inching closer to Afghanistan...
Neat!
help?
by Koen
hi. how do i make that thing work? need i first run generate matrix completely? and how do i enter keyword? btw: i speak dutch so pleae inform me.
UK wars?
by Tony ML
I was surprised not to see the UK in the war diagram. We are and have historically one of the most aggressive and defensive nations and I don't think there are many places in the world where our forces haven't fought - even if this is just the last century you have WWI and WWII, Korea, Vietnam, Cyprus, Central America (various), the Falklands, Yugoslavia, Iraq, Afghanistan, etc.
Interested to know the time scale used
Post a comment
| 
