Brownfield land registers

On 16th April 2017, the The Town and Country Planning (Brownfield Land Register) Regulations 2017 came into force. These stated that local planning authorities had to prepare and maintain a register of previously developed land - brownfield land - and then publish these as open data by 31st December 2017. With the start of 2018, and the data published nationwide, it was time to explore it.

What is brownfield land?

Land which is or was occupied by a permanent structure, including the curtilage of the developed land (although it should not be assumed that the whole of the curtilage should be developed) and any associated fixed surface infrastructure. This excludes: land that is or has been occupied by agricultural or forestry buildings; land that has been developed for minerals extraction or waste disposal by landfill purposes where provision for restoration has been made through development control procedures; land in built-up areas such as private residential gardens, parks, recreation grounds and allotments; and land that was previously-developed but where the remains of the permanent structure or fixed surface structure have blended into the landscape in the process of time.

Getting the data

Back in July, the Department for Communities and Local Government published guidance to help local planning authorities share their data.

Within West Yorkshire we have five local authorities. Leeds and Bradford published their datasets on Data Mill North, Calderdale on the Calderdale Data Works, and Kirklees and Wakefield put the files on their council websites.

Data wrangling

Thankfully, the brownfield land register format is defined so the datasets follow a common CSV file format. That makes life easier than it often is combining local government data from different places.

I wrote some perl code to grab the CSV files and convert them into GeoJSON as that's an easier format for me to put on a web map. But first I had to deal with coordinates. The register's specifications allow either WGS84, OSGB36, or ETRS89 as the coordinate system and, potentially, it can vary from record to record. Each council could choose whichever they like (and include an Easting/Northing if they want). I found a perl module that converts from the British National Grid to more usable latitudes and longitudes.

I adapted some code I was writing for another project, and created a map of West Yorkshire brownfield sites. Strangely, the ones from Kirklees were all missing from the view. Where were they? It turned out that the Kirklees dataset had put the latitudes in the GeoX column and the longitudes in the GeoY column in the CSV file. That meant all their points ended up north-west of the Seychelles in the Indian Ocean! I switched them around and all was well. At least with the Kirklees data.

Where's Kirklees gone?
Kirklees on holiday in the Indian Ocean

It turned out my coordinate conversion had a bug. I didn't find that out until we shared a version of the map on Twitter. Sydney Simpson at Bradford MDC spotted that Bradford's entire dataset had been shifted slightly to the east of where it should have been. Oh dear. I got on the case and found my bug. Sydney also spotted that the attributes for the Bradford locations seemed to be mixed up. As this conversation took place out in the open, Steve James at Bradford MDC realised that this was due to a problem with Bradford's CSV file. Within a few hours he had corrected it and re-published the data. This is a great example of one of the benefits of open data; when you have 'many eyes' looking at the data, in different ways, you can find the mistakes and improve the dataset.

Our little experiment to visualise new datasets has been pretty successful. We have an interactive map of West Yorkshire brownfield sites. And, thanks to responsive people in local councils, the published data have been improved too.

The final map