Data Journalists Engaging in Co-Innovation…
You may or may not have noticed that the Boundary Commission released their take on proposed parliamentary constituency boundaries today.
They could have released the data – as data – in the form of shape files that can be rendered at the click of a button in things like Google Maps… but they didn’t… [The one thing the Boundary Commission quango forgot to produce: a map] (There are issues with publishing the actual shapefiles, of course. For one thing, the boundaries may yet change – and if the original shapefiles are left hanging around, people may start to draw on these now incorrect sources of data once the boundaries are fixed. But that’s a minor issue…)
Instead, you have to download a series of hefty PDFs, one per region, to get a flavour of the boundary changes. Drawing a direct comparison with the current boundaries is not possible.
The make-up of the actual constituencies appears to based on their member wards, data which is provided in a series of spreadsheets, one per region, each containing several sheets describing the ward makeup of each new constituency for the counties in the corresponding region.
It didn’t take long for the data junkies to get on the case though. From my perspective, the first map I saw was on the Guardian Datastore, reusing work by University of Sheffield academic Alasdair Rae, apparently created using Google Fusion Tables (though I haven’t see a recipe published anywhere? Or a link to the KML file that I saw Guardian Datablog editor Simon Rogers/@smfrogers tweet about?)
[I knew I should have grabbed a screen shot of the original map...:-(]
It appears that Conrad Quilty-Harper (@coneee) over at the Telegraph then got on the case, and came up with a comparative map drawing on Rae’s work as published on the Datablog, showing the current boundaries compared to the proposed changes, and which ties the maps together so the zoom level and focus are matched across the maps (MPs’ constituencies: boundary changes mapped):
Interestingly, I was alerted to this map by Simon tweeting that he liked the Telegraph map so much, they’d reused the idea (and maybe even the code?) on the Guardian site. Here’s a snapshot of the conversation between these two data journalists over the course of the day (reverse chronological order):
Here’s the handshake…
I absolutely love this… and what’s more, it happened over the course of four or five hours, with a couple of technology/knowledge transfers along the way, as well as evolution in the way both news agencies communicated the information compared to the way the Boundary Commission released it. (If I was evil, I’d try to FOI the Boundary Commission to see how much time, effort and expense went into their communication effort around the proposed changes, and would then try to guesstimate how much the Guardian and Telegraph teams put into it as a comparison…)
At the time of writing (15.30), the BBC have no data driven take on this story…
And out of interest, I also wondered whether Sheffield U had a take…
Maybe not…
PS By the by, the DataDrivenJournalism.net website relaunched today. I’m honoured to be on the editorial board, along with @paulbradshaw @nicolaskb @mirkolorenz @smfrogers and @stiles, and looking forward to seeing how we can start to drive interest, engagement and skills development in, as well as analysis and (re)use of, and commentary on, public open data through the data journalism route…
PPS if you’re into data journalism, you may also be interested in GetTheData.org, a question and answer site in the model of Stack Overflow, with an emphasis on Q&A around how to find, access, and make use of open and public datasets.
Creating Thematic Maps Based on UK Constituency Boundaries in Google Fusion Tables
I don’t have time to chase this just now, but it could be handy… Over the last few months, several of Alasdair Rae (University of Sheffield) Google Fusion Tables generated maps have been appearing on the Guardian Datablog, including one today showing the UK’s new Parliamentay constituency boundaries.
Looking at Alasdair’s fusion table for English Indices of Deprivation 2010, we can see how it contains various output area codes as well as KML geometry shape files that can be used to draw the boundaries on map.
On the to do list, then, is to a set of fusion tables that we can use to generate maps from datatables containing particular sorts of output area code. Because it’s easy to join two fusion tables by a common column, we’d then have a Google Fusion Tables simple recipe for thematic maps:
1) get data containing output area or constituency codes;
2) join with the appropriate mapping fusion table to annotate original data with appropriate shape files;
3) generate map…
I wonder – have Alasdair or anyone from the Guardian Datablog/Datastore team already published such a tutorial?
PS Ah, here’s one example tutorial: Peter Aldhous: Thematic Maps with Google Fusion Tables [PDF]
PPS for constituency boundary shapefiles as KML see http://www.google.com/fusiontables/DataSource?dsrcid=1574396 or the Guardian Datastore’s http://www.google.com/fusiontables/exporttable?query=select+col0%3E%3E1+from+1474106+&o=kmllink&g=col0%3E%3E1
Using Google Spreadsheets as a Database Source for R
I couldn’t contain myself (other more pressing things to do, but…), so I just took a quick time out and a coffee to put together a quick and dirty R function that will let me run queries over Google spreadsheet data sources and essentially treat them as database tables (e.g. Using Google Spreadsheets as a Database with the Google Visualisation API Query Language).
Here’s the function:
library(RCurl)
gsqAPI = function(key,query,gid=0){ return( read.csv( paste( sep="",'http://spreadsheets.google.com/tq?', 'tqx=out:csv','&tq=', curlEscape(query), '&key=', key, '&gid=', gid) ) ) }
It requires the spreadsheet key value and a query; you can optionally provide a sheet number within the spreadsheet if the sheet you want to query is not the first one.
We can call the function as follows:
gsqAPI('tPfI0kerLllVLcQw7-P1FcQ','select * limit 3')
In that example, and by default, we run the query against the first sheet in the spreadsheet.
Alternatively, we can make a call like this, and run a query against sheet 3, for example:
tmpData=gsqAPI('0AmbQbL4Lrd61dDBfNEFqX1BGVDk0Mm1MNXFRUnBLNXc','select A,C where <= 10',3)
tmpData
The real question is, of course, could it be useful.. (or even OUseful?!)?
Here’s another example: a way of querying the Guardian Datastore list of spreadsheets:
gsqAPI('0AonYZs4MzlZbdFdJWGRKYnhvWlB4S25OVmZhN0Y3WHc','select * where A contains "crime" and B contains "href" order by C desc limit 10')
What that call does is run a query against the Guardian Datastore spreadsheet that lists all the other Guardian Datastore spreadsheets, and pulls out references to spreadsheets relating to “crime”.
The returned data is a bit messy and requires parsing to be properly useful.. but I haven’t started looking at string manipulation in R yet…(So my question is: given a dataframe with a column containing things like <a href=”http://example.com/whatever”>Some Page</a>, how would I extract columns containing http://example.com/whatever or Some Page fields?)
[UPDATE: as well as indexing a sheet by sheet number, you can index it by sheet name, but you'll probably need to tweak the function to look end with '&gid=', curlEscape(gid) so that things like spaces in the sheet name get handled properly I'm not sure about this now.. calling sheet by name works when accessing the "normal" Google spreadsheets application, but I'm not sure it does for the chart query language call??? ]
[If you haven't yet discovered R, it's an environment that was developed for doing stats... I use the RStudio environment to play with it. The more I use it (and I've only just started exploring what it can do), the more I think it provides a very powerful environment for working with data in quite a tangible way, not least for reshaping it and visualising it, let alone doing stats with in. (In fact, don't use the stats bit if you don't want to; it provides more than enough data mechanic tools to be going on with;-)]
PS By the by, I’m syndicating my Rstats tagged posts through the R-Bloggers site. If you’re at all interested in seeing what’s possible with R, I recommend you subscribe to R-Bloggers, or at least have a quick skim through some of the posts on there…
PPS The RSpatialTips post Accessing Google Spreadsheets from R has a couple of really handy tips for tidying up data pulled in from Google Spreadsheets; assuming the spreadsheetdata has been loaded into ssdata: a) tidy up column names using colnames(ssdata) <- c("my.Col.Name1","my.Col.Name2",...,"my.Col.NameN"); b) If a column returns numbers as non-numeric data (eg as a string "1,000") in cols 3 to 5, convert it to a numeric using something like: for (i in 3:5) ssdata[,i] <- as.numeric(gsub(",","",ssdata[,i])) [The last column can be identifed as ncol(ssdata) You can do a more aggessive conversion to numbers (assuming no decimal points) using gsub("[^0-9]“,”",ssdata[,i])]
PPPS via Revolutions blog, how to read the https file into R (unchecked):
require(RCurl) myCsv = getURL(httpsCSVurl) read.csv(textConnection(myCsv))
