From Media Design: Networked & Lens-Based wiki
Jump to navigation Jump to search

While making an application to retrieve pictures associated with websites on the basis of the geolocation, I realised that some big websites could come up with different locations on the basis of the url I typed.

'website': ' ',
'city': ' Menlo Park ',
'time_zone': ' America/Los_Angeles ',
'longitude ':' -122.1781 ',
'latitude': ' 37.459 ',
'country_code': ' US ',
'country_name': ' United States ',
'continent': ' NA '
'website': ' ',
'city': ' Kensington ',
'time_zone': ' America/New_York ',
'longitude ':' -80.948 ',
'latitude': ' 40.7212 ',
'country_code': ' US ',
'country_name': ' United States ',
'continent': ' NA '
'website': ' ',
'city': ' None ',
'time_zone': ' Europe/Dublin ',
'longitude ':' -6.2597 ',
'latitude': ' 53.3478 ',
'country_code': ' IE ',
'country_name': ' Ireland ',
'continent': ' EU '

I looked a bit more into facebook because I remembered some article I read about its European data centers, based in Ireland.

Three first accurate positions I could map, see:

Then I tested all the potential country extensions for facebook urls, using the Country code top-level domains list from wikipedia [1]. Some things are very interesting about this list. For instance the fact that .su (Soviet Union) is still in use. I exported all the relevant data for the existing domains as a json file for quick mapping.

A lot of interrogation marks there:

What was interesting to me was that, when ever a URL would get located on its own in a country outside the USA's borders, it would look like an anomaly, and mostly confirmed not belonging to Facebook. In other words, the geo-checking of facebook urls enabled me to identify the domain's owner identity.


I also exported the Libre Calc spreadsheet[2] as a csv for d3.js experiments.


Here it is :