Tuesday, April 28, 2009

MARAC: Flickr: An Image is Worth a Thousand Views

Flickr is an online photo sharing site that enables users to “tag” (i.e., provide descriptive and other information about) one’s images. In this great session, archivists working in a variety of settings highlighted its practical value to archives.

Barbara Natonson discussed a pilot project undertaken by the Library of Congress (LC), which wanted to learn how social tagging could help cultural institutions and participate in an online community. LC chose Flickr because of its popularity and because its application programming interface (API) facilitated batch loading of photos. LC’s experience should be of interest to many larger repositories.

LC determined at the outset that every image it placed on Flickr would be available via its own site and that it would post only those images that lacked known copyright restrictions. It then did some custom programming that made batch loading practical and made its copyright statement (developed in consultation with the U.S. Copyright Office) appear whenever one of its photos was displayed. It also purchased a Flickr Pro account ($24/year) that allowed it to add large numbers of images and view access statistics.

LC’s first photos went online in early 2008, and LC adds new photos on a weekly basis. As of mid-March 2009, LC’s Flickr images have gotten roughly 15 million views. Most of the traffic comes from Flickr itself, but some of it arrives via seach engines, which index user comments.

To date, approximately 4,500 users have commented on at least one LC image. However, 40 percent of the tags are supplied by a small group of people, and most of the comments concerning images accompanied by good descriptive information simply repeat that information or document emotional/aesthetic responses. Images that lack such information produce the informative tags and comments that LC seeks.

A core group of approximately 20 “power commenters” corrects place names, supplies additional descriptive information, does historical detective work, and incorporates LC images into Wikipedia, etc., entries. These commenters have also highlighted how places have changed over time; photos documenting changes and links to GoogleEarth accompany some of these discussions.

LC actively monitors its Flickr photosets for uncivil discourse, and staff incorporate user-supplied information into LC’s descriptive resources and periodically update Flickr users on LC’s work; this work takes about 15-20 hours per week, and staff rotate responsibility for it. LC has also started incorporating links to Flickr versions of its images into its online catalog.

Natonson noted that that there are some risks to Flickr (and, by extension, other Web 2.0 technologies):
  • Disrespect for collections -- Flickr privileges the individual image
  • Loss of meaning/contextual information -- LC links Flickr images to its descriptive information in an effort to remedy
  • Reduced revenue from photo sales
  • Undigitized collections are by definition excluded
However, there are also substantial benefits:
  • Collections are made more widely available
  • LC gets additional information about its collections
  • The visibility of specific photos is increased
  • LC’s Flickr presence helps win support for cultural heritage institutions
  • Users can mix past and present -- thus leading to a more informed world
Natonson also discussed The Commons, which Flickr has developed specifically for cultural heritage institutions wishing to provide access to images lacking known copyright restrictions and which tries to address the individual biases in Flickr’s existing terms of service. At present, 24 institutions are members of The Commons.

The other presenters highlighted how smaller repositories could make use of Flickr. Judy Silva discussed how the Slippery Rock University Archives, which uses CONTENTdm to manages its digital collections, has used Flickr to reach out to new audiences and experiment with Web 2.0 technology. Slippery Rock’s Flickr project, which made use of the university library’s existing Flickr account, centered on 41 digitized photographs taken by an alumnus during his time in service during the Second World War.

It took Silva one afternoon to load the images into Flickr and do some very basic (i.e., non-LC) tagging, and the rewards have been substantial: to date, Slippery Rock has gotten over 700 comments on these photographs, and one commenter forwarded the obituary of one of the people depicted in one of the images.

Owing to the success of this project, Silva is thinking of adding more recent images in an effort to get information from people who might Google themselves or their friends.

Malinda Triller was not able to come to Charleston, so her colleague Jim Gerenscer discussed how Dickinson College's Archives and Special Collections department, which also uses CONTENTdm, is using Flickr to publicize and obtain more information about its photographic holdings.

By design, the archives’ Flickr project was simple enough to be completed largely by undergraduates. The archivists identified images that lacked copyright restrictions, had appeal outside of the Dickinson community, and had basic contextual metadata, and students scanned the images and added them to Flickr.

Unlike LC and many other repositories, which create high-resolution master images in TIFF format and mount lower-resolution JPEG derivatives on Flickr and their own Web sites, Dickinson didn’t want to manage TIFF files. Students thus scanned the images in JPG format, at 100 dpi, and in grayscale or color as appropriate; in the future, the archives will rescan the images as needed. Project work is documented in a basic spreadsheet that contains the unique identifier, description (collection-derived or student-supplied), and title of each image.

To date, Dickinson’s Flickr photosets, which consist of images of an 1895 family trip to Europe, the 1893 Columbian Exposition, a school for Native American children, and construction of a major road in Alaska, have received 66,000 hits, which is a remarkable amount of exposure for a college archives; however, the archives recently learned that its Flickr account settings greatly limited the number of people who could comment upon the images, and it corrected this error a short time ago. The archives is really pleased with the project and is planning to add another set of images to Flickr.

I think that a lot of archivists are hesitant to embrace Flickr and other interactive Web technologies because they either don’t grasp their potential or fear that they’ll find themselves in the midst of a digital Wild West. This session highlights how repositories of varying sizes can use Web 2.0 technology without being consumed by it or losing physical or intellectual control of their holdings, and many of the attendees seemed really intrigued by these presentations. I suspect that The Commons will grow as a result of this session . . .

No comments: