If you do look at the genderized names, the ratio of female to male is
about 39% to 61%, which isn't far off the numbers that Roslyn got on her
survey. (42% to 58%, if I recall) That does mean that we have to assume
that the folks who didn't include a name are as likely to be male as
female, but I think this may be more than just pure coincidence.
On 1/29/13 1:28 PM, Eric Lease Morgan wrote:
> On Jan 29, 2013, at 10:21 AM, Eric Lease Morgan <[log in to unmask]> wrote:
>>> This service uses Freebase to determine the gender of names, and offers a
>>> JSON API: http://genderednames.freebaseapps.com/
>> Using the API and for a good time, I plan to extract the first names of subscribers from the Code4Lib mailing list and generate a pie chart illustrating the proportion of males to females. Sort of like digital humanities?
> I have done a bit of "genderizing" of names against the Code4Lib mailing list, and the results (I think) are rather inconclusive. Yes, most of the names are male, but fully one third of the names are "ungenderizable". See the blog posting -- http://bit.ly/WNp764
> Eric Lease Morgan
[log in to unmask] http://kcoyle.net