The Public Sector Geospatial Agreement: open identifiers are a step forward, but there’s still more to do to unlock the value of the UK’s geospatial data
Yesterday’s announcement by the Geospatial Commission of the new Public Sector Geospatial Agreement included details on some significant changes to the UK’s geospatial data infrastructure.
The new agreement is accompanied by further announcements about the outcomes of the “OS Open MasterMap” programme which was launched in 2018. The programme was intended to deliver on Government plans to open up more mapping data: a move we welcomed at the time. Over the course of the programme, we have participated in the Customer Advisory Group which has provided advice to Ordnance Survey on their implementation.
The programme has included the development of a new data portal for accessing OS open data and a new range of APIs and datasets. The data hub is still in beta but is due to be officially launched in the summer, at which point it will be the primary means of directly accessing OS data.
We are particularly pleased to see the release of open datasets for key identifiers, like UPRNs and street identifiers and a clearer policy on how derived datasets can be enriched with those identifiers.
We have repeatedly advocated for open identifiers, most recently in our paper on the UK’s geospatial data Infrastructure. Having authoritative identifiers for properties and streets will help to link together key datasets across the public and private sectors. For example, currently the Ministry for Housing and Local Government publish energy performance certificates for properties, and the Land Registry publishes price paid for properties, but joining those datasets together is difficult and time-consuming because the addresses they use tend not to be provided in the same form. If they both use UPRNs, this task becomes a lot easier. We get the most value from datasets when they can be easily combined.
The revision to derived data policies, to allow publication of property extent data created from OS MasterMap should support the publication of additional open data. The changes will enable other organisations, particularly the Land Registry and local government, to publish datasets that describe the geospatial extents of properties (which are useful for planning permissions) in addition to including identifiers for them.
Derived data restrictions stop public bodies from publishing or sharing some data they create using data from Ordnance Survey. Either because of the restrictions themselves or because those restrictions are complicated to understand so public bodies err on the side of caution. Relaxing some of these derived data restrictions enables public bodies, and particularly local authorities, to publish more open data, and to be more confident about making devolved decisions about how to strategically use data to achieve their goals.
We look forward to seeing these changes leading to the publication of additional open data. As a result of their proposed migration from being a trading fund and the removal of derived data restrictions, we hope to see the Land Registry announce changes to the licensing of its geospatial datasets such as the National Polygon Dataset.
Our recent survey showed local government struggles to publish data. So they are likely to need further support and guidance to enable the value of these changes to be felt more widely. We hope that the forthcoming Geospatial Data Strategy will explore strengthening their capacity, given so much useful geospatial data is stewarded by local authorities.
Over time, we expect to see a range of new products, services and analyses using geospatial data from local authorities and beyond, helping to digitise the planning system, understand the impacts of a changing climate, or help local authorities assess safer routes to schools
However, in some significant respects, today’s announcement falls short.
The new data hub provides a set of industry standard APIs and services for accessing a range of OS datasets and maps. Use of the APIs includes a free usage tier – up to £1,000 a month – that will allow some businesses to access OS data for free for the first time. The public APIs and free usage tier will help businesses to develop and launch a range of new products and services. The APIs are designed to smooth out the transition of prototyping, testing and launching a new product or service.
The full set of terms and conditions that govern the use of these new APIs and the data they make available, are not yet available. But the existing documentation explains that they can only be used for certain purposes. They can only be used to build public-facing products and services and not as a means to support internal uses and analysis within an organisation. You can’t use the APIs to develop internal applications, to support a customer service application. Or use them in back-office applications of machine-learning and AI. Or simply to improve internal management reporting and analysis.
Finally, there is little in the way of new open data. The original Open MasterMap project scope said that “Significantly more geospatial data will be fully open for businesses and developers to use, free and without restriction”. This is not the case. Free access to data is not the same as open data. Ordnance Survey still restricts what can be done with the geospatial data it stewards on behalf of the UK.
It is good that Ordnance Survey now offers APIs, reduced pricing, and a freemium pricing model that will be more attractive to developers. It is good for there to be competition around the provision of these services, and in particular an option that gives better and more accurate coverage of Great Britain than Google Maps. It is also good to see that there will be “new, richer data” available in future.
But the ability to unlock the originally projected £130m a year requires more organisations to use geospatial data, in more innovative ways, not just more competition within an existing market. That requires data to be linked up and combined in flexible ways and used within businesses to increase their productivity. It requires open data.
So while this is promising progress, we feel a lot more work is needed to strengthen our national geospatial data infrastructure. We are looking forward to the publication of the National Geospatial Strategy later this year with the hope that it will provide context for this announcement within a larger and more ambitious long term vision. We would like to see:
- A long term vision for Ordnance Survey, as the UK’s primary geospatial data institution, that incentivises OS to ensure that data is used as widely and well as possible, for broad economic and social benefits, rather than for short term returns. The UK needs a strong mapping agency and the Ordnance Survey needs to be sustainable, in perpetuity. Charging for services rather than data is a step in the right direction. But there’s a lot more to be done.
- Clarity that the new ten year PSGA agreement does not preclude a continued move towards a more open future for both existing datasets and the “new, richer data” referenced in the announcements.
- A resolution – or at least a direction of travel – to make address data open data. This data has an estimated value of £992m–£1.32bn each year to the UK economy, but was privatised in 2013 with Royal Mail. Address data is a key part of our national geospatial data infrastructure and should be available as open data.
- Support for businesses, particularly SMEs, to use geospatial data to help improve their performance and productivity, whether that data comes from Ordnance Survey or open sources such as OpenStreetMap.
- Support, guidance and investment in local government to help them release further geospatial datasets.