Norway’s Agency for Public Management and eGovernment, also known as Difi, launched the data.norge portal in 2011, following national directives on the openness of public data.
One of the main goals of data.norge is to foster better coordination, creativity and efficiency in relations between the public and private sectors in Norway. The Ministry for Government Administration, Reform and Church Affairs developed the NLOD, Norway’s Licence for Open Government Data, which is the licence recommended for use in the public sector and another important part of the national open data strategy. There is extensive information about the use of the licence and its exemptions in a dedicated section of the portal.
This open data project is different from other initiatives: rather than acting as a data store, data.norge works more as a cooperation facilitator. Its role is more proactive than simply being a middleman between other actors. Its main task is providing useful links for datasets anyone can reuse. There is also an ever-growing Datahotel for open data storage needs. This way, anyone can share its own datasets and also save them under the data.norge umbrella.
Policy Context
According to data.norge, the Freedom of Information Act is the most important policy for the site’s purposes. This law has been updated recently to implement the PSI Directive. The Freedom of Information Act makes it mandatory to provide data in machine-readable formats. In addition to this, data.norge complies with the Circular on Digitization, which is a “soft law” that applies to central and federal institutions. The Circular requires that all new and renewed public-sector systems must consider how the information they manage can be shared as open data, and design their systems for sharing, preferably through APIs. The government has also revised its “Guidelines for making public sector information available for reuse”. These new guidelines are heavily inspired by the Share-PSI project.
Description of target users and groups
Difi was commissioned by the Ministry of Local Government and Modernisation to provide an open data portal for Norway. The dataset descriptions are provided by the public agencies themselves.
The main goal of data.norge is to make it easy for users to discover, evaluate and use public open data. Following this philosophy, both the private and the public sector have been targeted as users. Right now, a common data catalogue for data inventories in the public sector is being developed. This data catalogue will be a major boost for open data in Norway, but for the moment data.norge acknowledges the need for a more intensive campaign of information and publicity. Many users still don’t know what is managed by the public sector in terms of open data, and what they can do with it. However, the future of data.norge will be the integration of both sides: the links section of the main site - which shows the datasets and the links to them and the Datahotel, which actually works as open data storage.
Description of the way to implement the initiative
The portal started in 2011 with a low-profile strategy, retrieving data from relatively small sources that could act as examples of data reuse for bigger providers. With that goal in mind, re-users have been good at registering their products on the portal and adjusting their code to promote open data. Another primary goal was to create value within both private and public sectors by changing existing processes and services into more data-driven workflows.
Since 2011, several hackathons have been held by Difi, the Norwegian Mapping Authority (Kartverket) and the Arts Council of Norway. These events have been popular catalysts for building ecosystems around public open data. Open data in Norway is now more focused on getting “difficult” data shared. There has been a noticeable shift in user needs: compared to the days when getting any open data was an achievement, users are now more interested on the sustainability of the data, its quality and its formats. This can be seen as a sign that open data is evolving and entering the mainstream.
The current version of the data.norge portal contains only datasets that are accessible to all users: these datasets have no restrictions beyond attribution. The portal policy allows only open licences that do not restrict commercial usage of data.
Goals and more goals
In data.norge it is mandatory to provide access to open public data in machine-readable format. This contributes to business development, a more effective public sector and a more transparent democracy, encouraging open licences. The portal does not host or store datasets; it contains only descriptions of datasets and links to their distributions. The reason behind this distributed system is to ensure that users get as close to the data sources as possible, fostering direct cooperation between users and providers. Difi did not intend data.norge to interpose itself as a middleman in established relationships between users and providers.
Since mid-2015, data.norge has natively supported DCAT-AP, which is the recommended standard for public organizations to describe their datasets. The next planned step is to harvest DCAT-AP catalogues directly from public organizations, and to improve the portal’s DCAT-AP output.
Technology solution
The current version of data.norge runs exclusively on Drupal 7. There is a proof of concept to test automated data harvesting of DCAT-AP catalogues. This project is built on the Elasticsearch/Logstash/Kibana stack; it uses Apache Jena Fuseki as the RDF triple store, with Drupal 7 as the front end. The harvester now being piloted is open source and available on GitHub (http://difi.github.io/dcat-harvester/). The code, developed by the Brønnøysund Register Centre, is also available. Data.norge is not linked data itself. It is mainly csv, json and xml data distributed through RESTful APIs.
A “data hotel” with open formats
Although no data is hosted by the portal, the data.norge promoters have figured out an alternative: a “data hotel” for publishers who want to copy their data from downloadable CSV files to a REST-style API service. The code for this is open sourced on GitHub.
To accomplish its goals for reuse and openness, data.norge pushes for open formats. This includes putting pressure on publishers who still register data on the portal that is only available in proprietary formats.
The data hotel is a free service through which data owners can publish their datasets without the need for to make investments or create infrastructures. While data.norge was created as a clean registry, without data, via the data hotel it makes data available from those providers who are willing to cooperate.
Technology choice: Open source softwareMain results, benefits and impacts
The portal does not require user registration, with the exception of public-sector staff who wish to provide datasets, and users who have developed apps or services using datasets provided by the portal. Perhaps because of this simplicity, its usage is increasing. For example, more and more private vendors are reusing information from the portal in their e-invoicing services: after the release of the companies register and the overview of the Norwegian recipients in OpenPEPPOL, private vendors and public agencies now have much better data quality for their invoicing processes, and a more solid foundation for data exchange between themselves.
According to statistics from data.norge, in 2016 the portal had 70,000 visits, with 282,000 page views. The site does not see a huge amount of traffic because users are pointed to the data sources, so they do not have to return repeatedly to the portal for the same information. This fits data.norge’s role as a “data discovery service”. In contrast, Difi’s data hotel had almost 140 million lookups in 2016, including for some of the datasets referenced in data.norge.
Most browsed datasets
As mentioned above, data.norge itself manages only descriptions of data, with pointers to the actual data sources, as long as these provide machine-readable data with open licences. The data itself is made available directly to users – either through APIs or as downloadable files – by the organisations responsible for managing it. The types and nature of the data vary: from cartographic data and 3D point clouds to simple tables and code lists. One of the most popular datasets is the Norwegian company register. Other examples are zip codes, data from the mapping authority, weather data, e-invoicing recipients, restaurant inspection reports, and digitised content from museums.
Return on investment
Return on investment: Not applicable / Not availableTrack record of sharing
The main data.norge community is on Twitter, and there are also regional Facebook groups to discuss open data.
The metadata is stored and shared with the European Data Portal. data.norge expects to have dataset descriptions from other data portals, like the National Geodata Portal, very soon.
Lessons learnt
The development cost for the first version of data.norge was roughly €55,000. Currently, Difi is running over 40 different portals, and open data resources are gathered and managed by internal developers. There is thus no dedicated budget for data.norge at the moment.
data.norge was created after a long research process. Open data portals from other countries, such as data.gov.uk, served as inspiration. The development of DCAT-AP has also affected data.norge to a large extent, and has led to the inception of a national profile (DCAT-AP-NO) with some extensions that are specific to Norway. Sweden’s Metasolutions, with its Entryscape Catalogue – an excellent tool that supports DCAT-AP – has also been an important resource for the data.norge developers.
Scope: National