Post: 22 December 2017
Yesterday DCLG’s Open Data Communities site published the Planning Inspectorate Appeals Dataset, as both linked data and a bulk download.
The initial release contains 90,914 records of Planning Inspectorate casework decisions over the past five years. The data is re-usable under the Open Government Licence (OGL).
In technical terms the dataset looks fine. It’s well prepared and presented and there’s good documentation.
But open data users should always do their due diligence on new releases. So here are my tiresome questions.
How much of the planning appeals dataset is personal data, and therefore not re-usable under the terms of the OGL?
The dataset contains names and addresses of appellants, many of whom seem to be individuals, as well as names of planning inspectors. Some of the other casework details could arguably also be biographically or economically relevant information about the appellants.
I’ve no doubt government is empowered to publish this information. Most of the casework records are also available through a search interface on the Planning Inspectorate site.
Presumably DCLG has considered the PARSOL guidance, though that really only contemplates online publication of personal data for the purposes of the planning process.
It’s certainly useful to have the data available as a bulk download. But if there’s no licence to re-use much of the dataset that means the legal scope for re-use is only the same as it would be if somebody had scraped the data. (I think.)
Re-use rights aside, how long can a business that downloads the planning appeals dataset legitimately retain any personal data it contains? DCLG has said it intends to publish the data on the basis of a rolling five-year period. How many businesses will just read “perpetual” in the licence and not even think about retention limits?
The address data raises another question. Bearing in mind the licensing debacle with Land Registry’s Price Paid Data, what assurance do we have that the planning appeals dataset does not contain data derived from Royal Mail’s Postcode Address File or Ordnance Survey’s AddressBase?
And the broader question:
How much responsibility do open data publishers have to alert re-users to potential issues like these when they publish datasets that are the product of complex processes?
Image credit: scaffold-1665165 by PIRO4D (CC0)