Digital Humanities

Networking

Posted by Emily Crockett

This week about Networks was interesting in light of one of the presentations at CenteringDH, Not One of Us Is Them: Diverse Proxy Phenomenology in Pompeii by David Frederick at the University of Arkansas. Specifically he worked with a method called “Network Topology.” When applying this method of networks to this project, a room is a node and a door is an edge with your people/information flow through those edges to different nodes. Clearly this is very different than the networks we looked at in class, but it’s a very similar concept. Additionally, in this case the model is used as a predictor as well as an indicator of the relationships between the rooms and doors. The particular way networks are being used in this project remind me of a quote from the Weingart article.

This is not to say it cannot be done, or even that it has not! People are clever, and network science is more flexible than some give it credit for. The important thing is either to be aware of what you are losing when you reduce your objects to one or a few types of nodes, or to change the methods of network science to fit your more complex data.

Demystifying Networks, Scott B. Weingart

Networks are definitely one of the more complicated topics we’ve tackled in this class because of the importance of math and statistics. In a sense it reminds me of the conversations we had early on in class about whether digital art historians need to have strong technical skills. Certainly at a basic level it’s not too difficult to connect pieces of information, but once you start looking at the different measures it becomes much more difficult. As an exercise I decided to see if I could find another gephi file to play around with. This led me to a file on a hero social network graph.

This is an overlook of the entire network that had 10,469 nodes and 178,115 edges. Clearly this is not entirely useful because it just looks like a big mess. You can see further towards the middle that there are larger nodes, but it’s not clear what those nodes are.

This is a close up of the graph that shows some of the bigger names in the network such as Thor of Dr. Strange. It also includes the heroes other names that they may go by, although not all of them include this, such as Captain America who’s “real name” is Steve Rogers. This points to another lesson we’ve touched on in this class of how important it is to clean your data and to make sure everything is consistent. While this dataset is interesting, it’s not particularly useful because it’s so ridiculously large. In fact, even though my computer is new within the last couple months and has some of the top of the line specs available, I could barely run this file, so other than being able to get the few screenshots I wasn’t really able to query it. While this isn’t necessarily related to art history it does indicate a need to cull data and really think about what’s important to your research question.

Another sort of network visualization that I think is very interesting is the “map of the internet” which shows the internet as a network diagram (which is essentially what it is). It shows you what the largest websites are (facebook, google, etc.). It’s valuable to be able to interact with a network as large as this, because from a 30,000 foot view, it’s just a bunch of mess. This particular network is also fascinating in just the sheer amount of websites that are registered. While this shows the general categories of the websites it does not include the actual edges or connections between websites because that would probably get much too unwieldy.

Another network that is more related to art history is this “Inventing Abstraction 1910-1925” from MoMA that is a reference back to Barr’s Cubism and Abstract Art diagram. This network shows the important relationships in Abstraction and who knew who in this world. The orange nodes show the people that had the most connections.

I could see how a network diagram could be important to my own work as I find myself often trying to think of who Veronese is connected with, and who he may have been in contact with to influence him. For example, I recently found out through research that Veronese spent time in Mantua and may have been connected to Giulio Romano. This could be a possible explanation for the more sexual nature of Veronese’s work. Although I don’t think it would ever be possible it would be interesting if one was able to create a bimodal network of artworks and hands. In this I mean if you knew that so and so’s workshop had 15 apprentices if you could map those apprentices to different art works. It’s clear that network diagrams could be very useful to many different questions in Art History, but once again we have to find the happy medium of being an Art Historian and a Technologist.

2 thoughts on “Networking

  1. Michelle

    Hey Emily, I really appreciate how many examples you provided in this post. It was interesting to see the screen shots of each in rapid succession to see just how many possibilities there are for incorporating networks into a variety of projects. I’m not sure I understand the example you referenced from the conference entirely. Then again, that confusion hints at just how much you can do with networks. One thing I realized from going through your examples is that I really think networks work best in small sizes. Just as your computer could barely deal with a large file, I think a user can only “compute” so much information at once. A huge network seems impressive, but can actually say very little in reality. One way that your examples get around this is by varying the way the information is presented. For example the color change in the MoMA example is helpful in quickly realizing what may be the most important information. As a final note, I think a network could be a great research tool for you! Even if not part of your final project it could be a way to literally connect the dots!

  2. Taylor Hunkins

    Hey Emily,
    Thank you for this detailed and informative post. The large graphed network of Heros got me thinking– how are we really supposed to interact with these types of visualizations. You note that zoomed out, this network is cluttered and a bit overwhelming to understand. At that scale, what new information do we gain? That the data set is large? Seems unproductive. Of course, we usually are able to zoom in and move around with such graphs, but even your magnified image seems confusing. Names are just randomly(?) placed in the network with some on top of circles (I don’t know what these suggest). What is the point of this network? What kind of information do we glean from it? It seems to me that this might be an example of a non-network network, which as you note, should remind us of Weingart’s advice not to appropriate methodology.
    On the flip side, I think that network theory/visualization could be a good tool for you. I am interested in some of your ideas about how a network might help us understand artist workshops and apprenticeships — definitely an appropriate use of this method!

Leave A Comment

css.php