When thinking about our session on ways to visualize the past through cultural heritage collections, I found that my ideas fell into two broad categories: how institutions might visualize individual objects and collections; and how researchers might want to use those objects and data for their own research. What follows in this post are some of my initial thoughts about what museums and individuals are doing now and challenges facing them.
Institutions often represent individual digital objects with a visual, like a photo with a caption and some metadata. Still, very few museums, libraries, or archives are doing much else to communicate visual details and meaning of physical objects online.
- Visualizing for Individual Objects:
- Scale: One of the biggest challenges in representing artifacts with digital images online is to illustrate scale. ArtsConnectEd is one of the few sites I know of that deals with scale by including dimensions and by using a visualization of a hand, elephant, building, to communicate scale: http://www.artsconnected.org/resource/21248/12/8.
- Movement: Two-dimensional images of 3D objects are generally all that is available, while dimensions, depth, and full examination of an object can be difficult to visualize as a user. QTVR is not used very often, even though this type of software that sews together multiple images of an object to create a 3D representation has been available for nearly 10 years. While possible, this process can be time consuming. Few history museums incorporate an inexpensive option of using short videos to can create a similar effect.
- Visualizing Object Networks
- Pieces of a Whole: Most history objects are related to other objects and embedded in stories about their production, exchanges, owners, uses, significance. Some of these things are part of a set, one of many related pieces–think of pieces of the USS Maine spread out across US—or panels from the Migration series, a factory whose pieces have gone in many directions. Though it is possible to re-connect disparate pieces online, this practice often is not done. One example is the King’s Kunstkammer, a partial reconstruction of the Royal Danish Kunstkammer (a large cabinet of curiosities) that currently exists in several physical collections held in several different museums. http://www.kunstkammer.dk/GBindex.shtml The design tried to mimic a cabinet and the idea of rooms, which actually is useful in visualizing the cabinet as assembled by the King.
- Geographical Movement: Representing the geographic life of an object, visually, is extremely challenging since most mapping software only allows an item to contain one location. Are there ways to store lat-long data that can be mapped to show how an object is created and migrates, such as the life of a t-shirt (field-to-factory analogy)? A researcher could create their own visualization of an object using their own map, but it would be nice if an institution could represent its objects connecting multiple data points on the same map.
As a researcher, I may want to use and interact with online collections to create some of these networks, for instance, but am often foiled by step one: finding collections and data online! I’ve found in my survey of US history museums that only 17 percent of those museums provide a searchable databases for users (level of data available varies by institution), while 37 percent offer no collections information at all (not even a finding aid or a summary). If you are interested in objects, there just isn’t much there.
Once I locate relevant collection objects online, often there is no way to harvest data, other than by copying and pasting into a database that I create. For example, the Arago site contains a huge online database of postal history resources, but I can’t get any of it out of Argo easily. A small number of museums offer APIs, and finding public OAI-PMH sets is challenging.
Once I, or an institution, create a database of object data, there are tools like ViewShare that then offers options for visualizing this object data that can then be shared. But, the task of formatting, entering/creating, normalizing data requires a lot of labor ahead of time, and even more so if a museum works at the object level or is creating unique networks for objects (recording individual videos, adding icons that symbolize scale, geolocating).
One option for institutions would to make their collections available and the data harvestable in some way. LAMs might find that researchers are more than willing to share their “curated” data back with the institution for others to see/use/learn.
Most of what I outlined above probably seems extremely obvious. The reason for sharing these thoughts was to keep in mind that while there are different ways to represent collections and to use collections data to formulate new scholarly questions, a lot of work is involved just getting to the stage of creating a visualization.
I am most interested in working with the group to see if we can figure out ways to visualize movement/migration of items (could apply to people too), and to represent networks of objects or an object’s network.
Great stuff, and totally non-obvious 🙂 All too often I think there is a mentality in these kinds of situations where people say things like “put it on a map” and take no time at all to articulate what exactly someone is trying to communicate, what role location plays exhibition, etc.
As far as movement, migration, and relationships over time these are two of the most polished visualizations I have come across.
Patterns of Change in Railroad Company Board Members, 1872-1894 : I like how relationships over time are on display here.
Journalism’s Voyage West. The timeline slider associated with dots of various sizes do a nice job telling a story that moves across the country.
Now those sorts of highly customized views aside, I remain curious about what it is that we would want to abstract from those sorts of things that we might be able to build into more generic tools. For example, now that Viewshare has histogram sliders in it, you can start to play around with zeroing in on points of time in relation to a map which has pins colored on it associated with some categorical information. That starts to get toward something closer to those high gloss visualizations. For example, play around with this quick draft I did with a data dump from the blue ridge parkway collection or this quick view I made of a set of postcards.
I would be curious if anyone in the group could help articulate a more general model for what we are interested in seeing and what kind of data we might be starting from?
Insightful observations as always. As I recall, we’ve had lovely conversations offline and on twitter about mobility of objects. And, I want to underscore your observations on thinking about geographical movement of objects. How do we allow multiple locations of objects. Indeed, museum objects are almost always removed from their original contexts–place, space, time, and even ‘neighbors’? Those who study the landscape, including Stewart Brand emphasize ‘flow’–the relation between structures and their surrounds. Once wrenched out of time/space most museum artifacts lose that flow. How can we recapture it? Of course their is no easy answer. And, capturing the ways artifacts and objects relate at different moments in time could be fascinating. That said, I am not sure *more* data is the problem. It actually might obscure, but that is another issue.
I think, generally, the type of visualizations done by the Stanford Spatial History project go over my head, but the growth of newspapers is actually one of the few that I like and think is understandable by a broad audience. The railroad company board members viz seems like it was a lot of work but it doesn’t really do much for me. On the other hand that type viz might make lots of sense to those researchers, and so it is helpful for them.
I didn’t realize that ViewShare now had histogram sliders. That definitely adds a nice layer. I’ll take a closer look at your sets, Trevor. Thanks!
I don’t think we can recapture that flow, necessarily. I think that’s what some of the AR programs try to do by making historical photographs appear in the contemporary landscape through the medium of our phone camera. But we might be able to virtually re-assemble objects in a way that makes us look differently at things when part of a bigger network/context, than when we see them in their normal de-contextualized state. This is one reason why I love the digital version of the Kuntskammer. The actual cabinet already re-contextualized natural and produced objects collected to exist w/in its chambers. Those objects were pulled out and redistributed to other places thus breaking that context and making another new one. And then they were virtually joined. Complicated!
At the risk of self-promotion, I wrote an article a few years ago that you might find useful here:
Garfinkel, Susan. “Recovering Performance for Vernacular Architecture Studies.” Perspectives in Vernacular Architecture 13, no. 2 (January 2006): 106-114.
http://www.jstor.org/stable/20355387
I’d like to think we could capture the flow, as Mark calls it–in fact, I’m sure we can. A big part of the problem is that the data doesn’t exist because neither historians, art historians, all manner of related scholars of the past and its artifacts, nor museum curators, archivists or similar cultural repository custodians, have been asking the questions that would generate the data Shiela is looking for. Our new access to digital methods for analysis and interpretation is allowing more of us to ask questions that we either didn’t ask before, or asked in a theoretical way with no real hope of implementation. So now we face the task of figuring out how to generate and apply that data in a meaningful way.
One of my own frustrations with the state of digital history right now is that we aren’t talking very much about the new sorts of research we need to do in order to gather the data–and then the standards for encoding and sharing that data–that would allow us to take full advantage of new and emerging tools. TEI is one practical, if highly complicated, example. What if there was a similar standard for creating history metadata for cultural objects?
Thanks, Susan for your comments and the link to your article.
I agree that we still are figuring out what questions we want to ask and what info we want to glean. But, I don’t think that we need another metadata or XML encoding schema. For now, we just need to make metadata available. For some of the networks I wish to see, linked open data might help create those connections but again, those objects and some metadata need to be available. If an object lived in 3 places and that is noted in its provenance, I could see that and extract that. It wouldn’t necessarily need to be in new fields. Museums are the worst offenders, in some ways, because most have collections management databases with tons of metadata available in-house. Moving online & extracting selected fields from those records to share is the challenge because of proprietary software used for CMSs, lack of staff expertise and time, or overall hesitation to expose possible gaps or sloppy input of metadata.
Pingback: Working Group on Visualizations for Cultural Heritage | Lot 49
Pingback: Content and Context: Visualizations for the Public? | Visualizing the Past
Pingback: » Content and Context: Visualizations for the Public? [bracket]
It would be wonderful to track the history of artifacts in the museum. History museums don’t think about this too much. But art museums do. Many art museums, and most auction houses, keep track of provenance in great detail, and the Nazi-era provenance project at the AAM has added a great deal of new data. Here’s what typical auction house data looks like (it’s from Christies, http://j.mp/zx4gk8):
Gorham and Beatrice Cross, Hingham, Massachusetts.
Charles B. Lalli, Mattapoisett, Massachusetts, acquired from the above, 1989.
Altman/Burke Fine Art, Inc., New York, acquired from the above, 1989.
William Vareika Fine Arts, Newport, Rhode Island, acquired from the above, 1990.
Acquired by the present owner from the above.
I don’t know how this is tracked in art museum databases – but it’s there, it would help with your question of the history of objects.
At my little museum now we have a touchscreen system installed that’s been created by Brown students. Called LADS, for Large Art Display on the Surface (http://www.cs.brown.edu/research/lads/) it includes three locations for each object. It’s been hard to know how to use these. Place made, place related to object, present location? Much harder (and less useful) than I would have thought.
Pingback: Content and Context: Visualizations for the Public? | Visualizing the … | Exhibits Site
Thanks for your suggestions, Steve.
Have you seen art museums that make that history/provenance available to the public either on a standard gallery label or in the digital record?
I hadn’t thought to look at Christie’s as an example, but I definitely will.
Since writing this post, I also have seen a couple of new possibilities that will allow for creating multiple date fields and associating places with each date, one is from an Omeka project in development. I’m sure I’ll more to report later in the Spring.
Sheila,
Museum folk do want to share – that is the honey that brought us. Persistent funding concerns involving keeping the doors open is a much than worries about gaps or sloppy input of data. At our museum, when we identify and open a portal to our collections, we will expect to gain new and substantial information relating to provenance. This will come – if the data can be found by search bots – which will be key our adopting a portal.
Greg,
I do know that there is a general urge to share, but there does seem to be some resistance. It was at my former institution and I see it with some of the folks with whom we work now.
Good luck with your efforts, and I look forward to hearing more.