On Tue, Mar 19, 2013 at 3:14 PM, Carmen Mitchell
<[log in to unmask]>wrote:
> Heh, well she works with teams of students and willing volunteers from
> native communities. The faculty member in question has been
> doing documentation and revitalization of endangered languages and has
> worked on language revitalization efforts with several communities,
> including the Oklahoma Kickapoo, the Jicarilla Apache, the Q'anjob'al Maya
> community in San Diego, and the Ixhil Maya community in Nebaj, El Quiché,
> Guatemala. She wants to make all her research and data public - and has
> permission to do so - but there needs to be some assessment and structure
> to this. It's a cool (if daunting) project.
Exiftool is definitely worth a look then for your images. Even cheap
digital cameras embed all kinds of metadata in photos, so in many cases
you'll be able to assign community, place, time, and even subject based on
a combination of exif analysis and knowledge of the photographer.
For example, let's suppose someone never sets the clock/date on the camera
and just takes a bunch of photos. The researcher will know something about
who took some of them and where. Since the metadata tells you which camera
they used, you can identify all the pictures that person took which will
probably give you some broad topical headings. If you have a photo with a
known date, you can estimate the others. And if you find one photo taken in
an identifiable place, you can reasonably deduce that all photos taken very
close in time were also taken at that place.
The more you learn about the photos, the more you can guess with a
reasonable degree of accuracy. Once you break your mass of 300,000 photos
into groups of a few thousand, you can hand them to students. If you dump
key metadata into a spreadsheet and let the students add access points and
suggest which should be deleted (i.e. the vast majority), you can impose
vocab and shove those access points directly into the metadata as well as
whatever system you're going to use.
Exiftool is a load of fun. Even though the metadata gets mangled when it
goes through different programs, there's often quite a bit of interesting
stuff that's left in. For example, if you ever wonder where someone is,
just wait for them a pic to a social media site. Even if the location isn't
obvious, a lot of phones embed GPS data right in the photo without the
knowledge of the person taking it. Of course, a lot of photos don't have
this info, but there's often interesting stuff to find.