I don’t often post explicitly about my experiences as a graduate student – but I want to let everyone know about the focus of my studies for the next four months. I am taking two courses that I hope will complement one another. One course is on Archival Access (description, MARC, DACS, EAD and theory). The other is on Information Visualization over in the Computer Science department.
My original hope was that in my big Information Visualization final project I might get the opportunity to work with some aspect of archives and/or digital records. I want to understand how to improve access and understanding of the rich resources in the structured digital records repositories in archives around the world. What has already happened just one week into the term is that I find myself cycling through multiple points of view as I do my readings.
How can we support interaction with archival records by taking advantage of the latest information visualization techniques and tools? We can make it easier to understand what records are in a repository – both analog and digital records. I have been imagining interactive visual representations of archives collections, time periods, areas of interest and so forth. When you visit an archives’ website – it can often be so hard to get your head around the materials they offer. I suspect that this is often the case even when you are standing in the same building as the collections. In my course on appraisal last term we talked a lot about examining the collections that were already present on the path to creating a collecting policy. I am optimistic about ways that visualizing this information could improve everyone’s understanding of what an archives contains, for archivists and researchers alike.
Once I get myself to stop those daydreams… I move on to the next set of daydreams. What about the products of these visual analytics tools? How do we captured interactive visualizations in archives? This seems like a greater challenge than the average static digital record (as if there really is such an animal as an ‘average’ digital record). I can see a future in which major government and business decisions are made based on the interpretation of such interactive data models, graphs and charts. Instead of needing just the ‘records’ – don’t we need a way to recreate the experience that the original user had when interacting with the records?
This (unsurprisingly) takes me back to the struggle of how to define exactly what a record is in the digital world. Is the record a still image of a final visualization? Can this actually capture the full impact of an interactive and possible 3D visualization? With information visualization being such a rich and dynamic field I feel that there is a good chance that the race to create new methods and tools will zoom far ahead of plans to preserve its products.
I think some of my class readings will take extra effort (and extra time) as my mind cycles through these ideas. I think that a lot of this will come out in my posts over the next four months. And I still have strong hopes for rallying a team in my InfoViz class to work on an archives related project.
Reading your post made me feel that you might be interested in Topicscape – software for organizing information in 3D ( http://www.topicscape.com ). Any constructive feedback you want to give us, even if it’s negative, would be most welcome.
I am doing some user interface work right now so I just dropped by your blog to revisit your SAA Poster presentation.
RE: Information Visualization. You might want to check out Robert Allen’s article: “Using Information Visualization to Support Access to Archival Records” (Journal of Archival Organization Vol.3(1) 2005.
I think the timeline concept is a pretty universal and powerful information visualization technique that can be applied for browsing collections of archival materials. Some of the commercial desktop photo collection packages are starting use this (e.g. Adobe Photoshop Album 2.0).
They basically display a vertical bar chart over a relative timeline (i.e. adjusts according to the period of time that is in scope). The bar chart instantly lets the user see those dates/periods where there are lots of pictures available for viewing. I think Album 2.0 even lets you move a slider over the chart to bring the pictures for that date/period into the active viewing area. Pretty handy.
RE: definition of record in light of digital/interactive visualizations.
In my research I’ve also been struggling with this issue and decided first to step away from the restricting concept of a record as evidence and then tried to strip down what information actually is and does. See the last few posts on my blog http://archivemati.ca.
Woa! I just found ‘Bungee View’ over at the Historic Pittsburgh website. Talk about information visualization!
I doubt whether this improves structured searching but it is certainly a cool toy for browsing an image collection.
What is a record in the digital world? “It depends … of course.” In the bad old days, when I was an unrepentant I.T. geek, it always seemed to me that if you had the information content available in a comprehensible and a retrievable form, everything else was just a matter of programming.
Mike Miller, who taught Electronic Records in the summer, talked about some of the issues with automating the case records of the FBI. Among other things, the question of migrating the data from old systems into the newer system, and then that system migrating into a newer system. Some of the earliest records did not indicate hair color, and they had not created the first database (think 80-byte EBCDIC punch card days) with a “hair color not available” field. Instead, they defined a one-byte color code for “Purple” and used that to indicate that no hair color was on the original record. After all — nobody has purple hair (in the 1960’s). You can guess the world of hurt that comes from mixing today’s hair styles with that data!
What’s intriguing is the fact that context matters, and it might matter a lot more further down the road. Did the users of the first FBI case file database thingy have the ability to do Boolean searches like you can do today? (Hell no — you probably had to have a programmer code a search pattern on punch cards, run a search against all the cards using that pattern. “AND” was probably implemented by running the search program in succession with two different search patterns. “OR” would require running it twice and then merge-sorting the card results). I think that the place where visualization will come into the picture will be to present a combination of the data that was available at that point in time, and a sense of how the data was retrieved, either organizationally or technically, but the people using the system at that time.
Is that part of “the record?” I think that depends on the purpose of the reference/research rather than the purpose of the record, per se (yes, I know — this is blasphemy to archivists). Do I want to know whether a snippet of intelligence about Muktata al B’zzar existed in the FBI database on the day before someone “airplaned” the Loch Ness Tower? That’s pretty straightforward database/data warehousing archives in theory. Do I want to know whether it’s plausible that the security advisor to President Gas could have found that information and responded to it, or that it stood out in the gallons of information sent to President Gas every day in his email alerts … that’s presentation and context. I think sometimes we spend a lot of time focusing on “what is the record” and not so much time thinking about preserving the context of how those records were used, or why. Why don’t I have a record group with the records retention and disposition schedules of the County Records Center? Doesn’t that say as much about the context of the (paper) records as do the surviving records in my repository? I think the same should be true of an electronic records archiving system.
Pingback:Visualizing Archival Collections - SpellboundBlog.com - ponderings of an archives student
Comments are closed.