Crowdsourcing, Archives and Museums—Two
Examples:
NARA's Citizen Archivists and the V&A's "Search the Collections"
The trickiest part of this assignment (to me, at
least) was finding crowdsourcing projects that were (a) current, and (b) worked
within the context of it; perhaps I was looking in the wrong places, but I
seemed to find many more examples of how to set up crowdsourcing projects for
your LAM institution than I did projects that were ongoing, did not require a
large investment of time, and were not primarily curatorial in nature. While the Museum of Fine Arts asking
museum goers to select the most representative Impressionist painting from its
collection through a process of elimination voting is one way to get the
community involved in the museum’s activities, I’m not convinced it really
counts as making an active contribution.
Another project that attracted a great deal of attention is the New York
Public Library’s “What’s on the Menu;” unfortunately, it appeared that most if
not all of the menus have already been transcribed, and for some unknown reason
I found navigating the site to help with proofreading (the only step left for
most, if not all, of the menus) all but impossible. I
finally settled on the following two projects:
the National Archives’ (NARA’s) Citizen Archivist Project for the archival
example, and the Victoria and Albert (V&A) Museum in London’s “Search the
Collections” Project for the museum example.
1. The
National Archives’ Citizen Archivist Project
From the front page of NARA’s web site at www.archives.gov, click on the link “Information For… Citizen
Archivists,” which brings you to http://www.archives.gov/citizen-archivist, the front page of the Citizen Archivist
Dashboard. The mission statement, or at
least part of it, is this: “One day…All of our records will be online. You can help make it happen,”[1]
and depending on your time, talents, and inclination, there are a number of
ways in which you can aid them in this pursuit.
Once you’ve set up an account, which is very easy, you can check out the
list of user projects, or “missions,” pick out one that seems appealing, and
tag, transcribe, subtitle videos, and upload and share documents and photos
(you’ll need to set up a Flickr account to do this last one). If, like me, you choose to do transcription,
you’ll find some helpful hints at https://www.archives.gov/citizen-archivist/transcribe/tips.html;
another link that perhaps applies more to those contributing tags and comments,
but which should be read by everyone, is the Citizen Contribution Policy at https://www.archives.gov/social-media/policies/tagging-policy.html,
which is clearly aimed at keeping the project free of spam and threatening,
harassing, or offensive language. One
more interesting aspect of the project is the History Hub, a six-month crowdsourcing
platform pilot program for American history where citizen archivists,
academics, archival professionals, amateur historians, and other likeminded
people can meet, work together, and share information—“[T]hink of it as a
one-stop shop for crowdsourcing information related to your research subject.”[2] NARA goes on to say:
“The National Archives aims to
connect with and better serve customers interested in the historic records we
hold. We are launching the History Hub
as a limited 6 month pilot project so that we can test its usefulness as a
crowdsourcing platform. We hope to apply
what we learn to a longer-term solution that can be used by federal government
agencies and other interested organizations looking to expand public
participation.”[3]
The interface for actually getting to the
missions once you’ve actually logged in via the “Transcribe” page isn’t
especially intuitive; you have to go to the “What’s New?” section of the menu,
scroll down and find the link referencing citizen archivists, and click on
it. (To add to the confusion, the page
says “Research Our Records” at the top, but it’s not the same “Research Our
Records” page that you get when you click on “Research” using the menu; it
really shouldn’t be this complicated.)
Once you’ve jumped through that hoop, however, it’s reasonably easy to
click a few more buttons to find a mission that looks interesting and get down
to work.
I decided to transcribe captions on National
Forest Photographs, went to “Historical Trees—California,” and immediately ran
into a problem: you don’t find out until you’re actually in a particular
mission and looking at an individual photograph whether or not its caption has
already been transcribed or if it has already been tagged unless you click on
the “View/Add Contributions” button; if it’s missing one or the other (usually
they seem to need tagging rather than transcribing), it’ll still be listed,
waiting to fool unsuspecting transcriptionists.
After going through this same routine with several other missions, and
becoming progressively more frustrated, I finally landed on “Historical
Buildings—Wisconsin,” and was finally able to transcribe several captions
successfully.
Overall,
I found the NARA Citizen Archivist Project interesting, and, once I worked through
the few roadblocks, the format was easy to use.
While one’s individual account keeps track of how many transcriptions,
tags, and comments one has made, broken down by month, year, and all time, I
was unable to find any such program-wide overview of the citizen archivists’
contributions to NARA’s mission of being the nation’s archive of record—surely it’s
been a significant contribution over the years, and particularly with the rise
on the online citizen archivists.
2. Victoria
and Albert (V&A) Museum Crowdsourcing Search the Collections
According
to its website, “[a]s the world’s leading museum of art and design, the V&A
enriches people’s lives by promoting the practice of design and increasing
knowledge, understanding and enjoyment of the designed world,”[4] so
it’s hardly surprising that a major component of the site’s search function
would be visual. After all, many people
might not remember the name of a given artist, but show them an image of a
particular work and they may well recognize it.
The new
version of the V&A’s “Search the Collection” function contains over 140,000
images of works made in almost every imaginable media, all reduced to small 2-D
images to be viewed on a computer, tablet, or even cell phone screen. Since the images in question were chosen
automatically, the feeling is that many of them may not be the best possible
view that could be displayed within certain parameters (all images must be cropped
into a square in order to fit the “Search the Collection” homepage format), and
that all of them should be re-examined with an eye toward selecting the ones
with the best details that are most likely to help the searcher find the
artwork in question. This, of course, is
where crowdsourcing comes in.
The first
step is to go to http://collections.vam.ac.uk/crowdsourcing,
set up an account, and log in; this brings you back to the main crowdsourcing
page, but with a slight difference: you
can now see at the bottom of the page a bar graph showing how many images have
been processed out of the total that require processing. Click on “Click here to begin,” and a new
page comes up with five slightly different “crops,” or views, of the same
image; usually the first one leaves the most space at the top of the image,
cutting off part of the bottom, whereas the fifth image is the reverse, cutting
off the top and leaving room at the bottom.
Select the one you think best shows off the word (it’s usually, but not
always, the one in the middle), and click on it. You then go to another screen with five different
versions of the same image, but this time the difference is the zoom
level: it ranges from a normal view to a
tight close-up. Again, select the one
you think best illustrates the image—I generally tried to strike a balance
between including as much of the original image as possible with showing a fair
amount of close-up detail—and click on it, and a different set of images of a
new item pop up.
In
addition to the images for clicking, there are two bar graphs near the top of
the page; one is labeled “Our progress,” and for some odd reason it doesn’t
seem to change (I suspect there may be a bug or badly written code at work
here), the other is “Your contribution,” and it shows both how many individual
objects you have processed, both in this particular session and overall. (Try as I might, I haven’t been able to find
anything similar for other volunteers, other than the general progress bar
chart on the main crowdsourcing page; it also doesn’t appear that there’s been
a great deal of work as of late, since I seem to be the only person to actually
process any images over the past week or so.)
While there is generally a very brief identification of the object in
the images (“Postcard”), if you click on a button labeled “Need to know more
about this object?” it will take you to another page which describes it in
considerable detail; I found this to be quite helpful in certain cases, especially
those involving photographs of buildings, as it helped give context to the
images. If you feel that none of the
images are quite right, you can click on “No good image” or “Skip this object,”
but I found that if you do that, it’s impossible to go back for a second shot
at it—it disappears, never to be seen (at least by you) ever again.
While in
many ways this was a much easier undertaking than transcribing photo captions
for NARA, I would have to argue that it’s not quite the child’s play it might
at first appear. For one thing, you need
to have a reasonably good eye for detail and sense of balance and
perspective. (One of the aspects of this
project that attracted me was that I like to think I do have a good visual
sense, and having a minor in Art History doesn’t hurt, either.) Deciding which parts of an image should be
highlighted, and which can safely be downplayed or even cropped out, also takes
a certain amount of consideration; I spent several minutes studying a caricature
of an actor which had been signed by the performer himself, debating whether
including his entire head or his entire autograph (which was rather large) was
most important, and cursing silently that the V&A had insisted on going
with a square format instead of a rectangular one. (I went with the head in the
end, since he would have looked rather foolish missing the top of it, and the
description of the item did include his name and the fact that it was
autographed.) Another item was a heavily
embroidered jacket—was the important part the jacket as a whole, or the embroidery,
and which should be highlighted? I tried
to strike a balance, and think I succeeded, but it’s definitely trickier than
it first appears.
3. Final
thoughts
Looking
at both projects, I see both projects being sustainable for some time—I would
say “until the material that needs processing runs out,” but somehow I don’t
see that happening for a very long time, in NARA’s case; it may happen
eventually for the V&A (although, given the current rate of progress, not
for quite a while), but since both institutions will doubtless continue to
acquire new items, they could keep going for quite a long time. Inviting civilians (as it were) to pitch in
and help with processing is certainly one way to help create closer ties with
their institutions; I know from firsthand experience that seeing a project succeed
that you played a role in, no matter how small, is a nice little ego
boost. My main concern here, at the risk
of being cynical, is that crowdsourcing could all too easily slide from “welcoming
people into the hallowed halls” to “hey, free labor to exploit!”; some
companies have already displayed leanings in that direction, and human nature
being what it is, I’m not naïve enough to believe that no one else would do likewise. (I’m looking at you, Jeff Bezos and your “Mechanical
Turk”…) I also didn’t see any indication
that anyone else is reviewing the crowdsourcers’ work; presumably someone is at
NARA, and I hope so at the V&A, but I didn’t see any specific references to
this, and while I know that proofreading normally takes a lot less time than
doing it all from scratch yourself (then again, you haven’t seen some of the
work I’ve seen), I can easily see a situation that ties in to my “free labor”
concern, where corners are cut for the sake of saving money and quality
eventually suffers. If we keep the
potential problems in mind, however, and work to avoid any such problems, then
I think there’s no reason that crowdsourcing couldn’t continue to be as
successful in the future as it is today.
[1] National
Archives. Retrieved from http://www.archives.gov
[2]
History Hub, National Archives.
Retrieved from http://www.archives.gov/citizen-archivist/history-hub/
[3] Ibid.
No comments:
Post a Comment