Visualising White Comments
I have spent the week fending off the raised eyebrows I get when I explain that I'm going to the annual SND Information Graphics conference, mainly because I am taking it as holiday and paying for myself.
SND is the Society of News Design. It is firmly based in media but with presentations including subjects such as airport design and wayfinding. It is one of the broadest design conferences around and largely bereft of people whose style outweighs their substance.
This has been an interest of mine since I was young and I was fortunate to work with designers who went to the conference when I led the News Online design team before I moved to TV commissioning.
What has this to do with anything?
Well, the BBC Two White season website includes Spectrum: a visualisation of comments from BBC News's Have Your Say (it's on the right hand side of the page).
We commissioned this in order to allow people to explore this complex debate more freely than they might in the conventional text format. As one of the BBC's most popular television services, BBC Two must make subjects like this accessible. But being BBC Two it must do it in a way that encourages discovery and serendipity.
As well as the TV programmes, we decided that the White Season would also be a good place to highlight some material from the archive and quality contributions in Video Nation.
Then the question of what we could do with comments came up.
At first, we were cautious.
I worked at BBC News Online for ten years and during that time became pretty familiar with the ideas and stories that ignite beyond just "interest" (these include immigration, Israel and the Middle East etc). We were determined that we would not show anything incendiary. So everything would need to be premoderated.
Some argue against the BBC doing this, with accusations of censorship. But the reality is that if we premoderate, the public won't miss out on spicy, talk-show, "robust" banter, but will be spared a morass of banal and offensive material.
Having decided that the main debate will be housed in Have Your Say for its editorial robustness, we then attempted to visualise what was happening. Much of the work started at database level. We interrogated past debates on this and similar subjects to assemble hundreds of featured adjectives.
With this list, we cross-referenced comments in HYS to label and segment those coming into the White Season debate. It was then an information and graphic design job to create an interface and functions that worked.
Information Visualisation is an aspect of digital media that is reaching out to mass audiences. The Facebook friend wheel is one well known example. We wanted the users to have more involvement, which is why we have enabled agreement/ disagreement on the user interface. The application display becomes richer the more people use it, as well as giving an overview of the main emotional themes.
One of the purposes was to enable more exploration of the data than normal. We seem to be averaging 30 clicks per user at the moment, so that is pretty good. We also wanted to provide different ways of seeing these data - the "Emotional Detail" is my favourite but there is also something very satisfying about the "Regional" view. We have been iterating this through the week in response to where people are clicking or not, as that kind of responsiveness is important in a temporal experience like this.
We are not the first to do this. BBC News has done some really nice work. The Election graphics and swingometer were a great success with audiences.
First of all, we're trying to get the experience right, balancing the compexity of functionality and data with ease of use. Secondly, we're looking for where this treatment is useful editorially. And like everything that we do, the post project review should help us decide where else we use this type of experience - as should your thoughts and comments on this post.
Max Gadney is Channel Editor, BBC Two & BBC Four, Multi-Platform, BBC Vision
© MMIX~RS~q~RS~~RS~z~RS~29~RS~)
Comments
Please visit http://simile.mit.edu/ to learn about creating accessible and interactive visualisations without requiring the use of Flash.
(No, this isn’t spam advertising, just a BBC licence fee payer that can’t use most of the interactive elements on the BBC sites)
John, in this case, flash is clearly required. In other BBC news cases it hasn't been, but if you look at the complexity of the animation here, you can see that it could never be done in html etc...
It looks really great. I'm not sure its terribly good at categorising the comments correctly, but it certainly does it with style :)
Perhaps you should investigate using JSON for passing data from the back-end, would be neater :)
I can see why the concept of Spectrum is appealing – Having a design background I bet the pitch was great [ ‘imagine swirling clouds of different coloured emotions’ ] – but it is far too simplistic a medium for conveying the nuances of a debate such as 'White'. I really don’t think it’s fit for this purpose, with this debate at least.
Were the ‘Spectrum’ emotion headings arrived at before the season started? If so then you started out with a preconceived idea of what the contributions would be. All the contributions listed under ‘Happiness’ for example would better be described as ‘Unhappiness’; what ‘Caring’ is supposed to mean is anyone’s guess, the contributions I saw under this heading would be better classed as Hurt or Confused.
Spectrum appears to be designed to a pre-set agenda: that the entire white working class are angry, hurt, fearful, confused and unhappy that no-one cares about them or their views (on immigration). This is stereotyping, inflammatory and doesn’t accurately reflect the wider picture. Never mind, for example, all the white working class women with mixed-race children, or middle class hostility expressed towards the WWC; that adds too much complexity into the picture.
More importantly I don’t think Spectrum reflects the full range of comments in the Have Your Say that Spectrum is supposed to visually represent.
Quite a few contributions posted in the actual ‘Are white working class people ignored in Britain?’ HYS don’t fit any of Spectrum’s categories. How, for example, would the following post (chosen simply because it was the first I came across on the 1st page of HYS at that time) be categorised?
“If certain sections of society displayed more pride and solidarity with their countrymen instead of constantly whingeing and chipping away at the freedoms and opportunities afforded them, then perhaps we could move forward together to create an even greater nation. Anyone, of any creed, colour or religion can be British, which should be a badge of honour and not an aspersion or thing to be deconstructed by 'little victories' against ones perceived 'enemies'. Out into the sun Britons!” Mark Dexter. 12 March, 2008, 10:36 GMT 10:36 UK.
The above doesn’t seem to fit into Anger, Fear, Hurt, Confusion, (Un)happiness or (Un)Caring, at least as defined on Spectrum.
Why no headings for Dispassionate, Analytical, Reflective, Acceptance, Compassion, Hope, Sympathy or Pride, Embarrassment? Or Paranoia come to that, several posts could easily come under that?
There is also an issue of how posts are categorised? Some could be placed under several headings, that decision-making process is filtered through someone else’s judgement and interpretation. I know many will be judgment calls that could easily have gone several ways.
Perhaps when making a post offer people the opportunity to categorise their own feelings from a pick list, with the additional option of ‘none of these’ or of entering one word they feel best describes their main emotional response to the debate. You could also ask if they consider themselves working or middle class (or other) etc. Let people catagorise themselves.
Effectively I’m suggesting a tagging system attached to individual posts that allows you to produce a more accurate, sophisticated visualisation. You could then present tag clouds of names linked to their comments etc.
I’m afraid I felt so strongly yesterday that Spectrum so inadequately and inaccurately reflected HYS opinion that I have submitted a complaint. All in the interests of continual evaluation and improvement of course.
I tried to use this thing and gave up as it didn't work very well. I couldn't work out how to use it and could not click on the balls as they were too fast.
Hi and thanks for your comment. Let me address what I think are the main points of your comment, some of which are addressed in the 'Info' panel at the top right of the page.
The most important issue to understand here is that the list of adjectives was taken from previous debates on similar subjects. The design and technical team spent a good deal of our development time making sure this data was at a decent standard.
We felt that by presenting the comments and connections in this non-statistical way that people would not seek any absolutes from this visualisation but would instead enjoy more serendipitous connections, unavalable elsewhere on the BBC.
I absolutely agree that future products like this can make use of alternative means of tagging or adding data to comments to ensure an even more rewarding experience. This is still early days in User Interface Development and I hope we will see more bold experiments in the market that develop these ideas.
The SIMILE Project is an excellent initiative, but partly as a result of being designed for non-expert users the sophistication of interfaces it supports are limited, and too limited for what we really need when trying to understand and interrogate complex data effectively.
Spectrum is a worth initiative: a rich dataset has been exploited, visualisation controls have been carefully considered, and - most significantly - it ties into another (established) view of the data: Have Your Say.
For me, the next challenges will be to tie comments into richer profiles of contributors (and link these to social network data so contributions can be filtered by people in one's network), and to find ways to inform people of relevant updates to the debate.
The SIMILE Project is an excellent initiative, but partly as a result of being designed for non-expert users the sophistication of interfaces it supports are limited, and too limited for what we really need when trying to understand and interrogate complex data effectively.
Spectrum is a worth initiative: a rich dataset has been exploited, visualisation controls have been carefully considered, and - most significantly - it ties into another (established) view of the data: Have Your Say.
For me, the next challenges will be to tie comments into richer profiles of contributors (and link these to social network data so contributions can be filtered by people in one's network), and to find ways to inform people of relevant updates to the debate.