On Tuesday the Department for Environment, Food & Rural Affairs (Defra) will announce the results of OpenDefra, its ambitious project to make at least 8,000 datasets freely available to the public over the course of a year.
OpenDefra was launched in late June 2015 by Environment Secretary Liz Truss. Yesterday was the deadline for Defra network organisations to meet their targets for publication of open datasets, as recorded on Data.gov.uk (DGU).
According to DGU, Defra has comfortably exceeded the 8,000 dataset target. The following nine Defra organisations have published a total of 9,573 datasets over the past year, of which 9,483 are re-usable as open data.
Defra’s own final tally may be slightly different. Defra has been cagey about exactly which organisations within its network were expected to contribute to OpenDefra or the precise targets they were given.
I’ve assumed that most smaller bodies, such as National Park authorities, were out of scope. However its disappointing not to see anything from the Drinking Water Inspectorate, the Veterinary Medicines Directorate, and (in particular) Kew.
What is a dataset, anyway?
9,483 datasets is clearly a lot of open data. But one of the minor controversies around OpenDefra is what exactly counts as a “dataset” for purposes of recording on DGU.
There are legitimate reasons why it might make sense to document a large body of data as a collection of datasets instead of as a single dataset. On DGU this is left to the discretion of individual publishers, so there is a wide range of practice.
By some small miracle Environment Agency has published 1,885 open datasets on DGU in the past year, out of a total 1,535 datasets (including unpublished and non-open) on its own National Dataset List.
I don’t think this is a deliberate attempt to “game” the OpenDefra target. Many EA datasets are technical and complex, and need more explication on DGU than they do on the simpler National Dataset List.
However it’s a matter of judgement whether the water body data released to support Water Framework Directive river basin management plans should have been catalogued on DGU down to the level of individual catchments (550+ datasets), or whether the Water Quality Archive needed a DGU record for each area and year (800+ datasets).