Sheffield Deprivation

I've been experimenting with deprivation data (again) and google maps (again). This time I've overlaid deprivation data for Sheffield on to google maps. I did this by using a fairly simple technique, which includes making sure the geographic projection is correct. I've described this before too.

You can pan/zoom on the embedded map below, but to see the full thing you'll need to click on 'View Larger Map'. Once you've done that you can see further information on the left hand side. You can turn layers on and off using the little tick boxes on the left and you can query the map by clicking it.

Not much else to say apart from take a look at the map below.


View Larger Map

A Nearest Neighbour Index of Multiple Deprivation

I've blogged before about my attempts to add a more spatial dimension to the existing Indices of Deprivation for England. In this post, I have decided to share the data for my calculations on a Nearest Neighbour Index of Multiple Deprivation, following the method described in this paper.

The image below shows the IMD2007 (left) and NNIMD2007 (right) for Birmngham, with red most deprived and blue least deprived.

I have put together a spreadsheet containing the original IMD data plus my NNIMD data for all 32,482 LSOAs in England. If you click on the links below you can access the data - you'll need to click the 'download' link to get the data.

NNIMD dataset - Excel 2003 version (.xls)

NNIMD dataset - Excel 2007 version (.xlsx)

Deprivation and Spatial Context - The NNIMD

Time for a short post on some of my more recent work. A lot of what I do deals with areas, deprivation and trying to understand the links between the two. Since I live in Liverpool and work in Manchester, I've got a good idea about how some of the most deprived (according to the Index of Multiple Deprivation) areas look, where they are and how spatially clustered they are. I also think that the way we identify which areas ought to be the focus of policy could do with some refinement.

So, here's what I've done. I took the 2004 IMD data and created what I like to call the NNIMD. That is, the Nearest Neighbour Index of Multiple Deprivation. I've written a paper on this for a journal (to appear around October 2009) but I decided in the meantime to update the work for the 2007 IMD and post on it here. The NNIMD takes the IMD scores for all neighbours of an area and averages them to give what you might call a 'spatial context' or 'neighbourhood' deprivation value for each lower layer super output area.

This process is repeated for the entire country (England) so that for each of the 32,482 lower layer super output areas we have a score which provides intelligence on the kind of neighbourhood it sits in, at least in terms of the IMD deprivation score. Yes, it's not perfect and yes there are issues with the definitions of 'neighbourhood' that we have to use, but it's a good start on the way to understanding local spatial context and the role this might play in neighbourhood outcomes. I plan to continue this work in the future, but for now here's a couple of graphics showing how Liverpool looks using the 2007 IMD (first map) and the 2007 NNIMD (second map). [Note: the colour spectrum goes from blue (least deprived) to red (most deprived).]

From Data to Knowledge using Google Maps?

About time for another post I think. This time about things I'm doing with data which I turn into useable information in the hope that it can generate new knowledge. Well, that's the basic idea. Specifically, I've been playing around with google mashup editor, kml and kmz files, and Index of Multiple Deprivation data for 2007. What I've been trying to do is work my way round methods to display IMD data on google maps so that anyone can view it, understand it and maybe even remember it.

One of the main problems with using IMD data is that it means using lots of polygons and when you try to load them in google maps they render very slowly and sometime you get a script error in your browser (IE and FF). This is something to do with the way it all works 'under the hood' but I've been trying to investigate ways to make rendering polygons quicker... We'll see. Here's an example I created quickly with google mashup editor for an area in North Liverpool:

http://jrf-googlemaps-test.googlemashups.com/ - see how polygons load slowly


I've been reading and listening and watching online and it seems once you get beyond about 100 polygons everything goes awry. The solution would appear to be image/tile overlays but then this has some limitations that go against what I'm trying to achieve (query clickability, for example). I know people at CASA have used similar approach with their Gmap creator (I've used this and it works very well) and I can understand why but I just wonder if there is another way...


Since I'm on the subject of CASA, everyone with an interest in GIS/spatial analysis should go and see MapTube. This is really part of a much wider project about spatial literacy - which I am all for! I've said enough today. Will try a less technical post next time.

The Index of Multiple Deprivation 2004 to 2007

The new Indices of Deprivation for 2007 were released very late in 2007 and since this is an area I'm particularly interested in, I thought I'd blog on the differences between 2004 and 2007.

I've been exploring IMD 2007 just to see how it compares with IMD 2004 since authors say that they are comparable (but they do of course acknowledge that the methodology is not exactly the same).

For the most acutely deprived wards, there's not too much change, but the most deprived LSOA out of all 32,482 has moved from north Liverpool to south Liverpool (from just beside Anfield, Liverpool FC's stadium to Speke, by Liverpool John Lennon Airport). The list of the most deprived 100 LSOAs (about 150,000 people in total) is still dominated by the North West, with 68 out of the most deprived 100 LSOAs in England. It is also interesting to note the co-location of football stadia and the most deprived LSOAs. This phenomenon is perhaps not that surprising but it is replicated across Great Britian in other indices of deprivation. It also highlights the massive gulf between rich and poor in a very real manner.

More interesting are those LSOAs which have seen a significant change in LSOA rank. Using the 10% cut-off for change in rank (i.e. a move of 3,248 or more places up or down the list) I did a quick bit of analysis just to see how areas might have changed. In total, 2,374 of England's 32,482 LSOAs saw a change in rank of more than or equal to a 10% shift in their IMD category. That's 7.3% of all LSOAs, which seems like quite a lot. Most of these were not among England's 10% most deprived, but a handful were. Seventeen LSOAs among the most deprived 10% in 2004 experienced an improvement in their IMD ranking of 10% or more between 2004 and 2007.

Conversely, there were 11 LSOAs which declined by 10% or more between 2004 and 2007 to place among England's 10% most deprived. Most were in London. Overall, it's interesting to note the changes in ranks of areas and how extreme these can be. I just question if there can really have been that much change in 3 years in some of these cases and if so, what has caused it (gentrification? housing market pressures?). Are those LSOAs which have shown large improvements big success stories or statistical anomalies?

Lots of interesting nuggets when we compare IMD 2004 and IMD 2007.