Accessibility links

We’ve all seen the infographics in the Guardian supplements or on the BBC News website: they’re colourful, engaging and communicate large amounts of information over relatively small press space. We understand them on a fundamental level.

Network visualisations on the other hand seem to alienate the majority and by their very nature are too complicated. Huge amounts of valuable data become beautiful amorphic pictures that fail to communicate important stories to a broader audience.


Upon looking at complex network visualisations one gets a sense of ‘geekiness’ for want of a better word, with a Zuckerberg/Silicon valley flavour to them (which could of course be because of the number of social network based visualisations that are out there due to the recent availability of data from these sources), academic elitism that boggles the minds of everyday people. Often I feel the aesthetic runs the risk of intimidating a more general audience because they are unintuitive, not engaging and too much like something you might see in a physics text book. The revelations can be lost in the context of the network as a whole because they can be overwhelming.


This is of course a real shame as the power of large, complex data sets can be invaluable, tend to come from extremely credible sources and are often the result of extensive and timely research. There is also a good chance that a large amount of information may be lost because it is not displayed in a way that allows it to be properly understood by a broader audience. Important stories are being observed and understood by a select few while the rest of us see only pretty pictures.


My role as the Connected Communities and Design intern here at the RSA is to develop a process to visualise large complex networks in a manner that is engaging, holistic and intuitive. These will be used to communicate stories within the network to a number of potential audiences; academics, policy makers, respondents to the surveys that the visualisations will seek to represent, the press and general public.


Moving forward I will aim to cater to all groups and will be exploring the middle ground between infographics and network visualisation and perhaps tying the two together. Data sets will be explored and visualised in network visualisation software (such as Gephi or UCINET) and then exported and rearranged/designed in a style that allows stories to be more easily extrapolated from the mass of data. The designs will need to vary depending on the audience and the story being told. Our visualisation process will allow for additional simplification and explanatory layers to be added to the initial complex picture.


Here is a network that has been visualised in Gephi using the Force Atlas 2 algorithm. With the use of a key one can begin to understand the network, identifying attribute categories which both yield the most connections and are most central to the network. Gephi allows visualisations to be exported as vector files, offering a level of zoom that makes intricate details within the visualisation clear to see. There is also a plug-in called ‘Seadragon Web Export’ which exports visualisations that are embeddable in any web browser allowing smooth network exploration using zoom and pan tools.  Click on the image to enlarge.


I will aim to strike a balance between displaying the true network to highlight the presence and scale of complexity (a vital part of understanding a network as a whole) with additional graphic design elements, icons, familiar tools of measure and small amounts of text to better tell specific stories within the data. It is my hope that insight and perspective will be easier to gain by displaying these graphics side by side with the complex network visualisations - not only by academics but by all of us alike.

As outlined at the start the nature of visualising large networks will bring with it inherent difficulties. Combined with these are the challenges associated with catering to multiple audiences. I come from a graphic design background and am trying to solve these problems in that vein. I intend to post ideas and mock-ups as I go and would gratefully welcome any advice, ideas or suggestions.




Be the first to write a comment

Please login to post a comment or reply.

Don't have an account? Click here to register.