Looking for worldwide country information.
February 16, 2010 1:52 PM Subscribe
I need to find the geographic center coordinates for each country worldwide. I can find county level data all over the place, but not country level. Not interested in postcode centroids, which seem to clutter up the results on searches significantly.
Response by poster: Thanks DA. I need higher precision, but it's a place to start.
posted by Nothing at 2:17 PM on February 16, 2010
posted by Nothing at 2:17 PM on February 16, 2010
Mathematica 7 has all sorts of data on countries, including something called "CenterCoordinates", which they claim gives the "latitude and longitude of geographical center". But it appears to be only to the nearest degree, which probably isn't accurate enough for you. I'm not sure how they did their calculations, but it might be possible to get more precision (but 30 seconds of looking didn't show me a way). Sample output:
In[90]:= CountryData["France", "CenterCoordinates"]
Out[90]= {46., 2.}
In[91]:= CountryData["USA", "CenterCoordinates"]
Out[91]= {38., -97.}
In[92]:= CountryData["Azerbaijan", "CenterCoordinates"]
Out[92]= {40.5, 47.5}
(I actually wonder how accurate you can get for this kind of calculation---would it make sense to have a center to an accuracy of tenths of a second, given you have to work from measurements that may not be that accurate? Are you integrating over the country area or something?---but this is way not my field. Anyway, if having Mathematica's output would be helpful, let me know.)
posted by leahwrenn at 2:32 PM on February 16, 2010
In[90]:= CountryData["France", "CenterCoordinates"]
Out[90]= {46., 2.}
In[91]:= CountryData["USA", "CenterCoordinates"]
Out[91]= {38., -97.}
In[92]:= CountryData["Azerbaijan", "CenterCoordinates"]
Out[92]= {40.5, 47.5}
(I actually wonder how accurate you can get for this kind of calculation---would it make sense to have a center to an accuracy of tenths of a second, given you have to work from measurements that may not be that accurate? Are you integrating over the country area or something?---but this is way not my field. Anyway, if having Mathematica's output would be helpful, let me know.)
posted by leahwrenn at 2:32 PM on February 16, 2010
Best answer: Here you go: Coordinates are in WGS 1984.
posted by Jimbob at 2:35 PM on February 16, 2010
posted by Jimbob at 2:35 PM on February 16, 2010
Hmm, it looks like my link, Mathematica's output, and Jimbob's link are all using the lower 48 only for the United States. The geographic center of the United States "is located in Butte County, South Dakota.... The geographic center of the conterminous United States (48 States and the District of Columbia) is located in Smith County, Kansas." All the answers above seem to have Kansas locations for the US. Something to watch out for and perhaps serve as a check.
posted by DevilsAdvocate at 2:48 PM on February 16, 2010
posted by DevilsAdvocate at 2:48 PM on February 16, 2010
Response by poster: Actually, calculated based on center of the largest landmass is probably best, or else France ends up in Spain (or the Atlantic? Just a guess.) due to French Guinea. Thanks a lot everyone.
posted by Nothing at 3:17 PM on February 16, 2010
posted by Nothing at 3:17 PM on February 16, 2010
Hmm, after a bit of looking, I'm skeptical of some of the coordinates in the list linked by Jimbob. For example: Chile, Vietnam.
posted by DevilsAdvocate at 3:33 PM on February 16, 2010 [1 favorite]
posted by DevilsAdvocate at 3:33 PM on February 16, 2010 [1 favorite]
Devil'sAdvocate raises some fair points - this is a non-trivial thing to do, as you have to deal with (a) points for long, weird shaped countries falling outside the country boundary and (b) countries that have multiple parts (like the US).
I've recalculated using an alternative method, and the results are here. Chile looks more sensible, and the point for the US now seems to lie more in the South Dakota area.
posted by Jimbob at 3:49 PM on February 16, 2010
I've recalculated using an alternative method, and the results are here. Chile looks more sensible, and the point for the US now seems to lie more in the South Dakota area.
posted by Jimbob at 3:49 PM on February 16, 2010
Response by poster: Thanks a lot Jimbob. Do you mind if I ask what you are using the generate this?
I've seen several sources use center coordinates and label coordinates, where label coordinates are forced inside the boundaries of the largest polygon in the group. But these were for counties, not countries.
posted by Nothing at 5:46 PM on February 16, 2010
I've seen several sources use center coordinates and label coordinates, where label coordinates are forced inside the boundaries of the largest polygon in the group. But these were for counties, not countries.
posted by Nothing at 5:46 PM on February 16, 2010
I'm doing this in ArcGIS 9.2 using the FeatureToPoints tool, based on the world_adm0 data set - I can't tell you the original source of this shapefile, except it seems to be used by everyone and anyone, so I assume it's public domain.
posted by Jimbob at 5:55 PM on February 16, 2010
posted by Jimbob at 5:55 PM on February 16, 2010
Thanks, Jimbob--I was curious to see if there were any countries for which the centroid lay outside the country itself (at least for countries which are mostly contiguous), and from your latest set it appears that the centroid of Vietnam is in Laos.
posted by DevilsAdvocate at 8:15 PM on February 17, 2010
posted by DevilsAdvocate at 8:15 PM on February 17, 2010
« Older Really love your peaches wanna shake your tree | How does a time machine really work? Newer »
This thread is closed to new comments.
posted by DevilsAdvocate at 2:09 PM on February 16, 2010