This week, I write of nodes, edges, centrality, bimodes, and more – these new concepts for me are particularly confusing, to be honest. So, I’d take my analysis with a teaspoon of salt. My previous experience with networks include “what’s your wifi network password?” and “social networking is a strange form of delicious evil,” so needless to say, I’m a new-bie. This doesn’t mean, though, that I can’t find ways to engage with network theory and digital humanities projects, so hold on to your hats, folks, it’s going to be a bumpy ride with no in-flight wi-fi.
To start, I’ll cover some key concepts – the first of which is a node. In networks, nodes are points of information which often are shown connecting to each other. Nodes connected to more often have more centrality. Centrality increases when more instances of it occur in the data. This does not necessarily mean that centrality means something is more important, though – just that it has more connections to other nodes.
The visual connections between nodes are called edges. Contrary to what I thought, edges aren’t just connections – they themselves are a type of data. For example, if you were onnecting Tolstoy (node type 1 – author) to War and Peace (node type 2 – book), the edge would mean “is author of”. But, if you were connecting Audrey Hepburn (node 1 – actor) with War and Peace (Film)(node 2 – film), the edge would mean “stars in.” Connecting two types of information, stars to movies, authors and books, etc, means a network is bimodial – the “bi” prefix meaning 2 kinds of nodes. This understanding of networks as many living pieces is new to me – and helps me see them as livingorganisms as opposed to confusing, static visuals.
A challenge with these visuals is embracing the living aspect of them without sacrificing complexity – something that the nature of nodes necessitates, notes Weingart in his article, “Demystifying Networks.” I recommend this article to other in the humanities grappling with the uncertain nature of our data and the certainty that forms of analysis like these require.
This week, as I worked with both Palladio and Gephi, my concerns about relaying uncertainty and my desire to capture the action colored my interactions with both the software and the networking examples I analyzed.
Palladio is certainly fun to use. It creates connections similarly to Tableau, though it seems helpful with the mapping and timeline features and graphs, and is more intuitive than Tableau, which is also a little less refined looking. You cannot recolor or resize in Palladio, which is a feature I’d have liked to see. You can download the results of your network analysis – what you are downloading is a screenshot of part of your network. Since you don’t have to have an account, it is user friendly at first, access-wise, but you cannot save it for later, which is a considerable downfault, in my opinion.
Gephi is far more complicated than Palladio, but you have more control. To begin, open Gephi, open graph file, and get to work! When you make the graph, you’ll have circles and lines and can have different colors. These nodes and edges are moveable, depending on the theme and the tightness you give the overall network. It is extremely not user friendly – to change appearance, for example, you must click partition, modulatory class, palletes, generate new pallete, and move even further from there – and none of those titles are terms I, as a non-expert, would be able to associate with changing pink to blue. Not that I could do that anyway – the color variations do not offer a wide variety.
Gephi, as I have said is not intuitive, but great for editing specifics. You can spread things out more in layout – the Reingold layout makes the network into a pretty ring and spreads it out, which is actually quite beautiful. But getting your data into Gephi is a whole process and very complicated, so if you want your pretty data ring, be prepared to work hard for it.
The one redeemer for Gephi’s difficult interface is the online user manual – they have a lot of information in the online manual that is very specific and useful for anyone hoping to use Gephi. You can save Gephi projects, which is a step up from Palladio, but to present in a presenation at a conference you’d likely need to screenshot the static network, which is the same process as Palladio. As it is not based online, it’s unlikely that you’d be able to embed it as an interactive network within a presentation.