Research & development for the social web
The Newsography Experiment
The complaint of regional bias in news reporting is, more often than not, concerned with the media’s fixation on the capital’s institutions and populace. What basis is there for this argument? And is it possible to determine the extent of regional under representation whilst proving London enjoys near constant attention from the UK media? The Newsography project was initiated to answer these questions.
Having launched Channelography, a web service that analyses programme captions to provide new ways of exploring content from the BBC, we decided to look at another prolific producer of content – the Guardian. Parsing a large portion of the UK content from 2009 using Muddy (our text mining and analysis service) provided us with data that we could use to test the theory of London news bias. To do this, we commissioned info-graphics that use the data from Muddy to help tell the stories. The visualisation shows the looks at number of news stories per million people for each city and interestingly, the results challenged the initial proposition with London appearing to be under represented given it’s population size whilst cities such as Oxford, Cambridge and Brighton feature much more prominently than larger cities.
Exploring the data.
Looking for more patterns we instructed Muddy to find the entities that commonly occurred with each city and whilst a few notable ones were discovered such as the ‘Taliban’ commonly occurring with Birmingham, further work is needed to gain more meaningful insights from this process. Here are a few of the more common co-occurances :
| City | Entity | Mentions |
|---|---|---|
| London | Barack Obama | 230 |
| Edinborough | Lockerbie | 129 |
| London | Scotland Yard | 220 |
| Birmingham | Taliban | 63 |
| Birmingham | Mexico | 60 |
| Glasgow | Michael Martin | 59 |
| Glasgow | Lockerbie | 58 |
| Edinburgh | Libyan | 56 |
| Liverpool | Steven Gerrard | 51 |
At Rattle we’re interested in the analysis of news stories in their narrative and trends and whilst the Newsography exercise has ended it has sparked a few leads. This year we’re going to continue to work on discovering other interesting ways of looking at news in the meantime take a moment to download the pdf’s to see how your city ranked in 2009’s news reports.
7 Comments
Aden Davies
February 22, 2010What splendid visualisations…How about if you had access to the home locations of the reporters? Any correlation in the amount of times they wrote about their home towns for instance?
Journopig
February 24, 2010Is there any correlation between the apparent bias towards Oxford and Cambridge, and the amount of new research that is published by those universities (and publicised by their press offices)? I would imagine that would help explain the number of mentions of these cities.
PS Might want to look at your spelling of Edinburgh, BTW…
andrew glover
February 24, 2010Interesting study. Am notorious pedant so can you spell Edinburgh correctly please?
Thanks
James Boardwell
February 24, 2010@Aden – yeah, anecdotally we’ve heard from journalists that Brighton’s coverage is skewed because of the disproportionate number of them that live there. Not easy to get reporter address data though
@journopig – having looked at a sample of the stories the Unis of Oxford and Cambridge (and their output which is often deemed to be of particular significance) does skew the data, although we’d argue that this is a legitimate skew (thanks for sp tip btw!).
Perhaps we could open up the data for people to interrogate and play with so you can come up with your own theories on why the news is skewed?
links for 2010-02-24 « Onlinejournalismtest's Blog
February 24, 2010[...] Rattle » The Newsography Experiment Having launched Channelography, a web service that analyses programme captions to provide new ways of exploring content from the BBC, we decided to look at another prolific producer of content – the Guardian. Parsing a large portion of the UK content from 2009 using Muddy (our text mining and analysis service) provided us with data that we could use to test the theory of London news bias. To do this, we commissioned info-graphics that use the data from Muddy to help tell the stories. The visualisation shows the looks at number of news stories per million people for each city and interestingly, the results challenged the initial proposition with London appearing to be under represented given it’s population size whilst cities such as Oxford, Cambridge and Brighton feature much more prominently than larger cities. (tags: guardian visualisation bias regionalnews) [...]
Rob Lee
February 24, 2010@journopig, @andrew apologies for the lapse in spelling, now corrected.
Rattle » Week #1253
February 26, 2010[...] finally got round to talking about ‘Newsography‘ our mini project looking at how fair news coverage is across the UK. It uses our concept [...]