Post: 2 February 2014
As far as I can tell only four local authorities in the UK currently publish their school catchment areas as open data: Hampshire, Cambridgeshire, Brighton and Hove, and Lichfield.
Update: I missed Nottingham (HT SK53onOSM), and I gather Trafford MBC are finalising their dataset.
A number of other councils provide maps of their school catchment areas as PDF files, or via “interactive” web mapping applications, or both.
However generally speaking where maps of school catchment areas are available to the public at all they are buried in the admission prospectuses of local authorities, in formats that don’t encourage analysis or reuse of the information.
Do school catchment areas really exist?
It may seem surprising that school catchment areas have not been picked up as a priority on the UK open data agenda, given the widespread perception that the middle classes are preoccupied with getting their children into the best schools.
Surely there are numerous markets for this information? Why hasn’t the Department for Education released a national dataset of school catchment areas?
The short answer is that school catchment areas are not used in every part of the country. Admission authorities may operate any admission arrangements they choose, provided they comply with applicable legislation and the School Admissions Code.
The admission authority may be the local council, or it may be the governing body of an individual school or some other entity, so there is a range of different approaches.
Admission policies typically take into account any special needs of individual children and where their siblings go to school, but geography is always a factor to some extent; as a matter of common sense most children attend a school within a reasonable distance of where they live.
Catchment areas are a particular tool available to admission authorities when individual schools are over-subscribed, or for planning purposes when they want to make sure school resources keep up with demographic projections.
However even when an admission authority needs to prioritise school intake based on pupil’s home address it may simply use proximity to the school instead of a catchment area. The difference in approach is not always transparent, so it may seem to parents that a catchment area exists even if it has not been defined as such by the admission authority.
Even where formal catchment areas are used, boundaries are likely to be revised regularly to make sure that school capacity is in line with demographic changes and new housing developments. School catchment areas are not a form of administrative geography in the sense we usually understand that concept.
Do we need school catchment areas as open data?
In principle if a council is administering an admissions policy that includes catchment areas I would like to see the council publish those boundaries as open data, in KML and/or shapefile formats. Once the dataset has been created it seems a missed opportunity not to make it available for reuse.
There is clearly also some interest in central publication of locally held school catchment datasets, and the Open Data User Group is reportedly developing a business case.
However as there is no national policy requiring use of school catchment areas, and as the catchment areas that do exist are meaningful only within the context of particular admissions policies, there is only limited value in collating a national dataset from existing local datasets.
On the other hand I think there would be substantial value in open data release of a national dataset that documents the ‘effective’ catchments of individual schools.
Effective school catchment areas: a national dataset
The Department for Education should be able to produce such a dataset relatively easily, as a statistical extract from the National Pupil Database.
The core content of the dataset would be a count of each school’s population broken down by either Output Area (OA) or Lower Layer Super Output Area (LSOA). Where the counts are large enough to be non-disclosive, it may also be practical to include key stages and some basic demographic information.
OA and LSOA boundaries are Census geography and already available as open data. This means the above statistical dataset could be used to generate a boundary for the effective catchment area of any school.
Ideally the dataset should be complemented by information on which primary schools feed into which secondary schools.
This DfE dataset would only provide a current and retrospective picture of school catchment areas. However I think it would be of broad interest to parents making school choices.
More importantly, when combined with other available geodemographic information the dataset would support comparative analysis of school populations in fields such as health, crime, insurance and risk management, and social policy.