One domain where data visualisation is helping interpret data and incorporate narratives is journalism. In the UK we are familiar with the Guardian’s data journalism and data blog and the BBC is making more use of data journalism and equipping journalists with the skills and tools to interpret and depict a growing volume of data.
As part of exploring this trend we were tasked with watching the Journalism in the Age of Data video report by Geoff McGhee from Stanford University examining the use of data visualisation in story telling. This interactive video report features input from several practitioners in the field and provides an insight into their work. It consists of a series of videos but also related information and further links to follow which was a bit like finding an opening up a treasure chest of data visualisation inspiration such as the beautiful Flickr Flow which plots a colour wheel for the 12 months of the year based on extracting colour proportions in Flickr photographs. The result is beautiful art from seasonal colour.
Data Visualisation in Practice
In the second section various research and news organisations discusses their experience and approaches using data to illustrate new stories such as:
- How Different Groups Spend Their Day from the New York Times
- Road Deaths in Great Britain 1999-2010 from the BBC
- Historical Hurricane Tracker from MSNBC
Some contributors criticised data visualisation as being aesthetic but meaningless. The video featured debates on whether visualisations are ‘deep’ or ’shallow’ and whether visualisations are created for designers, with an eye on a prize, rather than their readers. Others disagreed arguing that, unless intended as art provoking questions and curiosity rather than journalism aiming to inform, data visualisations that look great but don’t provide any insight are bad examples of the craft. Should data visualisation be a category of artistic expression or journalism infographic or both? Are readers data literate enough to understand data visualisations? Certainly for data journalism, the telling of stories through data, that narrative element is crucial.
Data Visualisation as Narrative
Chapter 3 explores the narrative economy of data visualisation and how it borrows and extends genres from other media whilst adding some narrative innovations via a 2010 research paper Narrative Visualization: Telling Stories with Data by Edward Segel and Jeffrey Heer (formerly at Stanford now at the University of Washington Interactive Data Lab). This paper identifies seven genres of narrative visualisation:
- Annotated Chart
- Partitioned Poster
- Flow Chart
- Comic Strip
- Slide Show
Practitioners estimate you only have about 4 to 5 seconds to grab the attention of a viewer. For large and complex data sets then chunking the data and opening up detail through interaction can be really powerful. The paper identifies a three typical structures on a spectrum of interaction from author-driven to reader-driven:
- Martini Glass
- Interactive Slideshow
Other innovation include multi-sensory approaches to data visualisation by combining media and narrative techniques from multiple genres. This ‘data vizeo’ genre includes film, video and animation (motion graphics) that is rooted in data. GOOD magazine not only has some fantastic infographics but are experimenting with short data based videos via their YouTube channel.
Putting the Data into Visualisation
So far most of the report emphasised the visualisation aspect but any good data visualisation requires a quality data set. Chapter 5 discusses the ubiquity and pervasiveness of realtime data about events and poses the question are we reaching a time where no event will exist without an accompanying data visualisation? Nicholas Feltron’s Annual Reports are one of the best known and sophisticated examples of visualising personal data from an individual’s life stream.
Chapter 6 covers the demands of exploratory data analysis lifting the lid on visualisation to expose the data sources and handling beneath the surface. Also debated are the qualities of data particular completeness and openness. For example, the UK Government has a new data.gov.uk site filled with data and apps to use in support of open data.
“You have to learn how to sketch with data” – Amanda Cox
Expanding the Toolkit
Further tools and technologies were discussed in addition to the Processing language we are using. Many organisations use Adobe Flash for creating visualisations. They also mentioned the Flare library and Protovis. These latter examples look a little dated now, certainly Protovis is no longer actively developed and has been superseded by D3.js. The big advantage of D3.js is it is based on the standard web stack rather than a proprietary framework:
“D3 is not a new graphical representation. Unlike Processing, Raphaël, or Protovis, the vocabulary of marks comes directly from web standards: HTML, SVG and CSS. For example, you can create SVG elements using D3 and style them with external stylesheets.” – D3.js
Also covered were tools for non-programmers such as spreadsheets, Google Charts, Tableau Public and two tools we tried out in DITA: Wordle and Many Eyes. There are also cautionary tales about the rise and fall of free tools such as Swivel. The key message is there is no one way to do data visualisation and when selecting subjects, design approaches and tools the best thing is to play to your strengths not fashion. It all comes back to the story you are trying to tell.