The high-value datasets set out in the Open Data Directive have now finally been specified in a proposal for an implementing act. The datasets themselves are listed in the Annex. Everyone has the opportunity to comment on it until 21 June 2022. Six categories of data are covered by the directive and the act.
- geospatial
- earth observation and environment
- meteorological
- statistics
- companies and company ownership
- mobility
Further acts covering more categories may follow over time (and perhaps clearer regulation for the existing categories).
Lantmäteriet (the Swedish mapping, cadastral and land registration authority) carried out a government assignment to prepare the Government for the negotiations on this act within the Commission, and I was somewhat involved in that work.
The conclusion was that in principle all data held by the public sector within these categories is of high value, otherwise it would not have been collected. The value that could be realised by publishing it as open data was estimated at between SEK 10 and 21 billion per year for Sweden. And THAT only took into account a subset of all the data affected by the act, and not all the types of value that could arise either.
A deeper analysis is needed to see whether what the act covers corresponds to what Lantmäteriet based its calculations on, but at first glance the act is quite extensive for most categories, with the exception of statistics, where it is rather detailed and the selection is limited.
My first reactions
Here are my first reactions after a quick read of the act.
The act points to ”CC-BY, equivalent or more open” as the licence. It would be desirable to have wording that more clearly favours CC0 as the choice of licence.
There are requirements on formats, but they only cover file or delivery formats, not content formats. The same data from all EU countries that is not comparable does not lead to a common European data market, which is the aim of the act. Work on data spaces is under way that may change this, but the act contains no requirements.
There are no requirements for persistent identifiers. This is a great pity, as they would have created good conditions for use.
For statistics, the act is very limited in what is to be published, and there are no general requirements for statistics to be published as open data. Something as basic as election results, for example, does not appear to be covered.
Company registers will become completely open, probably with an exception for sole traders whose personal identity number serves as their organisation number. It will be interesting to see how the Government handles this and how it affects the remit and funding of, for example, Bolagsverket (the Swedish Companies Registration Office).