Erevnametrics: Visualizations

At the moment, all the RO builder gives as an output is RDF, which isn’t all that exciting for many people. To remedy this, several visualizations will be created, so that the users of this tool can have something to look at that isn’t just a mass of triples. Here I’ll list the ideas that have popped up so far for visualizations.

Idea one: a (customisable) web page. The idea is that it will create a web page (or small group of pages, depending on size) based on the RDF given. As an example, for the sorts of objects I’ve been working with mostly, there would be a section for projects associated with the research, a section for associated people, one for papers, etc. Each of these sections would be internally linked to each other, that is, a paper’s representation could list all the authors of the paper, and each one of these would be a link to the main section of that author in the page. The idea is that it should be somewhat customisable too, at least to change the order of the sections, delete unessecary ones, that sort of thing; as such I envision some kind of edit mode for the page, possibly after authentication of some kind, that would be seen like an overview of the page – each individual section and subsection would be hilighted, and they ought to be drag/droppable into different positions. There would also be a icon on each block to allow it’s deletion. A couple of default stylesheets will be provided, but as the structure of the document should be consistant, it’s easy enough for anyone to create their own/change a pre-existing one.

Idea two: a graph, differing from the usual RDF graphs you see, in that this would try to keep it simple – at around the same level of complexity to the end user as the RO builder is, that is, only the significant relationships relating objects, and hopefully arranged in some way that makes sense (all people in one place, all papers in another). There are several ways this could be accomplished, two that I’ve been looking at include the RDF Graph Visualizer, part of the OAT framework, and using graphviz to generate a SVG graph which could then have all kinds of fun going on with something like raphaeljs. Because the graph would be visualizing things a reduced level of detail, for people who want to see the detail, clicking on an element could take you to either a website-representation of the element, or just a graphite dump of it.

Idea three: similar to PeoplePivot, something that attempts to visualize the research through time. This would probably take the form of a webpage, with a bar at the bottom that controls how you move through the research chronologically, that is, you would be able to select (on a slider) a range, and a date, and the page would display the content that was published within that time period, as well as be linked to the other content that may not have been published/released/had anything significant happen to it at that time. It could also work quite well as a graph, but rather that showing/hiding elements, it would rather be hilighting them or making them otherwise stand out, with some other way of showing which events are the reason why the resource ought to be hilighted, else the graph would end up not looking very meaningful.

1 Comment on “Erevnametrics: Visualizations

  1. The time idea is interesting. I had only considered research objects as a whole view of the research, but highlighting it over by time would enable the viewer to see what order things were done.

    How does it affect the contents of the RO? Do all of them have time information? For example, could you add the time ranges for which a particular person was involved in the research object? It could be interesting to see when a person joins or a leaves a project, or even more so when a project completely hands over from one person to another!

Leave a Reply

Your email address will not be published. Required fields are marked *

*