Quite a lot of new instructional material is appearing online about techniques of data journalism. If journalists are being outwitted by government and corporate number-crunchers it’s because we haven’t stayed on track with digital analytical skills. Most of us are aware of this but not sure how to go about learning them.
Here’s the link to a very good, short (48 page), free PDF that touches all the bases, from how to use Excel to how to visualise the data you’ve unpacked and hand it to readers in a digestible form. http://www.tcij.org/sites/default/files/u4/Data%20Journalism%20Book.pdf
DATA JOURNALISM, by Elena Egawhary and Cynthia O’Murchu (Centre for Investigative Journalism and Financial Times).
Also new is Paul Bradshaw’s “Scraping for Journalists” to be found here:
SCRAPING FOR JOURNALISTS by Paul Bradshaw (The Guardian and Birmingham City University). This isn’t free (you can donate $15 or he would like $20!) but it’s certainly worth having. It condenses a lot of what appears on Bradshaw’s regular blog.
A thing to keep in mind about doing data is that the techniques are meaningless unless you have questions. Data is means to an end, not an end in itself! Ask questions and then go seeking the data that might provide answers. Data may also raise questions directly – as when crime figures from the police seem suspect, especially when you know that more crime has happened in your street than sees the light of official data. As Bradshaw says, there are scores of sources, and you can “scrape” them for the figures that reveal hidden patterns.
An example of data visualisation that prompts questions about population densities in one’s own country is this map of Australia . Australia was once described by the novelist Patrick White as a fringe of leaves surrounding apparently nothing. To the Aborigines it’s not nothing (they want it all back) but the map does make the point very clearly that a comfortable life is only to be found on the edges.
Pictures like this provoke questions. Everyone is living on the edge, literally. What’s missing from the Australian economy and the psyche of the people?
The question for us could be how to represent the population spread of South Africa or Namibia or Somalia (then: where to find the stats and, using Excel, and group them under geographical regions). It’s obvious that the economy of SA is focused on Gauteng and a few other cities, so to take that one step further, comparisons of income and quality of life between urban and rural dwellers can be represented on maps. How to find the data? Scrape it, as Bradshaw says – from official stats, private repositories (if you can get access) and humanitarian aid research sources. It’s a job to compile it all, and time-consuming too. Bradshaw has plenty of advice on how to do it on deadline.
I’m working on population and voter maps for South Africa’s upcoming elections and plan to get something on this blog soon. In reporting workshops on election coverage for the IAJ and others, I plan to introduce data mapping as key technique that helps to objectify the news. Where, say, rural voters have disproportionate representation (much more or much less than they should have if equity applied) we could show the resulting distortion of the Parliamentary seat count. President Zuma is said to have boosted the ANC’s total count in the 2009 election with masses of votes from Zulu supporters in KwaZul/Natal: again this can be reflected as provincial “sizes” of parties on a map.
Here’s another take on population mapping:
It’s a colourful way of representing populations in the “World of 7 billion”, where the size of the region reflects its population – eg India is huge but southern Africa rather small by comparison, and Mexico seems as big as the whole of the United States, while Canada is just a sliver.
Have fun with data mapping! – Graeme Addison @ngmuntutools