Investigating the Geography of Open Source Software through Github

"We collected our data through Github’s public API, which offers the same data as available on the Github’s website but presents it in a structured format for simpler processing. The data were collected from May to July of 2010"

"The data collection followed a recursive procedure. We started with a single account, belonging to one of Github’s founders. We then identified accounts connected to this user, then looked for accounts connected to the newly found ones, repeating this procedure until we achieved closure. New accounts were identified through the four kinds of connections mentioned in the previous section: (1) those that follow accounts collected earlier, (2) those followed by the accounts collected earlier, (3) those whose repositories were being “watched” by accounts collected earlier, (4) those who had made code contributions to the repositories watched by accounts collected earlier."