Among yesterday’s front pages was a data visualisation which, at first glance, was one of the most effective I’ve ever seen: the Independent had made an infographic showing yesterday’s £6bn budget cuts in context - as a fraction of a debt mountain.
Then I looked closer - and something’s very, very wrong.
Can you tell what it […]
Help Me Investigate is - quite simply - fantastic. It’s a site designed to help people collaberate on investigations, usually based around questions. An early question was “on which Birmingham streets are the most parking tickets issued?”. And it’s been answered in style.
Heather Brooke used Freedom of Information to get hold of the full record of tickets issued in the last year. I posted a few (rambling) thoughts on the figures on the site, while this guy did a rather more thorough job. With graphs. The Birmingham Post then ran the story in full - and to its credit, cited HMI thoroughly.
All great stuff. And now the site’s out of beta, users are no longer restricted to asking questiond about Birmingham, which must come as a relief to users outside the West Midlands.
The only problem is, I’m worried it might just leave me jobless. Which would be, as they say, a bit of a bugger.
Data is slowly, painstakingly, starting to catch on in UK journalism. Computer Assisted Reporting, as data journalism is clunkily and unhelpfully dubbed, is huge in the US, with almost every paper having a CAR editor and many having full CAR teams. In the UK, the situation’s different - outside the web team and tech desk, even reporters able to do a bit of excel analysis are few and far between.
The Centre of Investigative Journalism has been spearheading a move to change that. Each year it flies over US CAR experts to train up UK journos, who are also being encouraged to start training up their newsrooms, and current journalism students in using data to produce stories (not mashups, not cool data, not spreadsheets, though all are nice - the emphasis is on stories).
Thanks to their efforts, and some damn good work priming the pump from some early freelance adopters - including (yet again) Heather Brooke - a bit of a market is starting to emerge. Stories based on FOI and data analysis will sell, just. Slowly it’s becoming possible to practise this newer brand of investigative journalism. With a bit of momentum, it could yet hit the newsrooms proper, as compared to chequebook journalism, this is cheap as chips.
And here’s where Help Me Investigate, despite its brilliance, gives me chills. Some of the best data journalism is incredibly complex, fraught with legal issues, and inordinately time-consuming. Some downright middling data journalism comes close. This article on police compensation was a CAR story.
The most complex bit of data-analysis in that story was simply working out rates per 100,000. But gathering FOI data from 43 police forces, in different formats, getting them into one sheet, cleaning the data, and working out reliable population estimates was both lengthy and dull. Then the real work started: contacting each force before publication to give chance to respond - then dealing with each force coming up with a string of excuses and (much lower) revised numbers. This relatively straightforward story ended up taking well over 50 man hours.
Most don’t. But if papers can get stories that look and feel like “investigations” from sites like HMI very cheaply - even free - the rationale for hiring data journalists or buying in their stories gets weaker, especially given the newsroom climate of constant cost-cutting.
Projects like the Birmingham car parking tickets investigation are great targets for collaberation, work everyone should be glad is being done, and perfectly suited to sites like HMI. The more complex stuff is likely less so. It would be a crying shame if a scheme like HMI led to less of this work being done.
There is an alternative school of thought that leads to a virtuous circle - data journalists can work with HMI on some investigations and keep a steady stream of compartively straightforward stories flowing. The journalists most eager to help out on the site will have the inside track to publish, and in turn will also have more of their own time to work on the complex stuff that takes full-time workers.
If CAR were more established in the UK, and newsroom culture wasn’t what it is, I’d tend immediately towards this latter happy option. But at present, the majority of the UK’s data journos do their stories as freelancers - and losing this sort of low-hanging fruit both hits income (and given the “news mix”) makes the pitching battle harder.
It would be a crying shame if the brilliantly intentioned - and rather nicely executed - social journalism project that is HMI actually ended up stifling a fledgling journalistic field. Maybe I’m far too negative. But it does seem the risk is there.
With thanks to the twittered contributions and back-and-forth from:
@paulbradshaw and @podnosh - two of the site’s founders
and @rasga and @coneee
Thoughts gratefully received…
Gathered the below information for the excellent - and newly out of beta - helpmeinvestigate.com . Unfortunately, a gremlin in the system stopped this post appearing over there. So it’s here.
Right, here’s a round up of what’s out there through FOI so far - this is a frequently-FOI’d topic. Here’s what’re released - and what they’ve refused to release to previous requesters.
First up:
http://downloads.bbc.co.uk/foi/classes/disclosure_logs/rfi20080397_percentage_of_london_households_without_tv_licence.pdf
In which we learn:
Nationally, 2% of households don’t own a TV and an estimated 5.1% dodge their licenses. Only 4.3% of english household evade, while 11.1% in Northern Ireland dodge - interesting quirk, no?
The beeb refuses to release more detailed regional breakdowns in case it affects compliance rates (they suggest high levels of evaders might persuade others to dodge).
HOWEVER - An Information Commision ruling from this year (March) ordered the BBC to release some regional data - number of households with TV license and number of prosecution notices issued in each region - to the requester within 72 days. Presumably they’ve done that, but they haven’t put it on the log. Worth asking their press office/FOI officer to send to us too?
http://downloads.bbc.co.uk/foi/classes/disclosure_logs/rfi20080243_tv_licence_evasion_2007.pdf
Here we see how many evaders were caught in 2006 - 413,000 or so - but no mention of fines, prosecutions or otherwise. This is refused under the extremely obstructive grounds of being already published (Section 21(1)). Under Section 16 - duty to assist - they should link to, or attach, that information, and haven’t. Bad form, BBC. Number of enforcement officers is refused on the grounds of prejudicing prosecution of crime (s31 I think).
This one is fairly helpful:
http://downloads.bbc.co.uk/foi/classes/disclosure_logs/rfi20080008_enforcement_and_collection_costs.pdf
This is the enforcement costs etc as a percentage of income over a period of years.
Some figures from a pressure group here: http://www.itsmummy.com/tv_scabs_1.htm
There is also a rather comprehensive set of 141 FOI requests (including requests for detection equipment etc) here: http://www.onebillionpageviews.org/foi.html
An ICO ruling on the detection equipment is here: http://www.onebillionpageviews.org/downloads/FS50154106.pdf
Hope all that is helpful. I am trying like the blazes to find this home office stuff on the number of successful prosecutions (I can find reposts, which is useful but unreliable, but not the originals.)
If the prosecution numbers were accessible, we could maybe couple those with the numbers from the ICO appeal mentioned earlier, and try to infer whether areas have higher prosecutions due to higher delinquency or (over?) zealous enforcement in some regions.
