Linked Open Data: A vision for interlinking and sharing open data

From PDF to linked open data – that’s the vision of many open data enthusiasts when it comes to preparing and sharing open data. By providing examples of practical applications, the speakers at the 11^th Municipal Open-Data Network Meeting on August 17, 2023 described how this vision can be realized.

As Mario Wiedemann, senior project manager in the Bertelsmann Stiftung’s Data for Society project, pointed out at the start: the topic is not entirely new. While preparing for the online meeting, he came across presentations on linked open data given as early as 2012 that have lost none of their relevance or significance. “While we were looking for speakers, however, it became clear that not much has happened in practice at the municipal level,” he said.

The first speaker was Mila Frerichs, who provided an introduction to linked open data. He has been working in this area for 10 years – first as a volunteer and now also as a consultant. Frerichs expressed the goal concisely, saying, “We don’t want to share documents, but data. Although municipal authorities are increasingly publishing data, they often do so using a manual process that ties up resources and requires a lot of discussion first.”  

One example he gave was the creation of the school development plan for the district of Neuss in the Rhineland, which uses five different data sources for its forecasts, sources that make data available in different formats. “If you’re lucky, then they’re mostly CSV files and you don’t have to figure them out again each time,” he said. “The easiest thing, however, would be if you could create a query once for the different data sources and it would work each time, even if new data were added. That would be possible with linked open data.”

But how can this innovative concept even be defined? “The term ‘linked open data’ refers to a way of finding data, describing data, connecting data and using data,” Frerichs said. The following steps are necessary, he explained, to publish linked open data:

1. Model data
2. Use existing vocabularies
3. Create new vocabularies, if necessary
4. Assign URIs
5. Publish URIs with descriptions
6. Convert data to RDF
7. Publish data

Frerichs also presented the classification using five stars developed by Tim Berners-Lee. One star is given for shared PDF documents, five stars for linked open data. In the latter case, URIs from other sources are also used, so that data from different providers can be linked together. Users access one provider’s data and find new data from another linked source.

For this to work, data must be standardized. “We have to describe data in detail by modelling and standardizing them,” Frerichs explained. “The best way to do this is to sit down with representatives from other municipalities and consider together how new data can be created and standardized. In doing so, as many existing ontologies as possible should be used.” One example he cited is schema.org.

Frerichs conceded that the biggest challenge is creating, standardizing and relying on ontologies. Yet the advantages are obvious, he said, since a common foundation increases the quality of the data and the possibility of linking them.

By now, the approximately 90 participants, who attended the network meeting despite its taking place in summer, had a solid overview of linked open data. A lively discussion was already underway in the chat when Richard Hunkel, head of Open Data & Digital Projects at the German National Tourist Board, presented the Knowledge Graph for German tourism.

In his 20-minute presentation, Hunkel made a connection between artificial intelligence and linked open data. He showed how Internet use will change significantly in the coming years – from keyword-based use to dialogue-based use. Artificial intelligence will continue to gain in importance, he said, and applications like ChatGPT understand structured data better. “In contrast, the texts, brochures and documents that many tourism professionals still work with cannot be used in these contexts,” he explained.

When someone plans a city getaway or a vacation in a specific region, Hunkel said, they currently have to gather information from a number of sources – about the journey, the hotel, concert tickets, opening times of museums, the weather. Ideally, all the relevant information should be available in one place. “Our approach is to gather all the content about a particular destination in a single content hub, i.e. the Knowledge Graph, so that an application can be created from it,” he explained.

According to Hunkel, the project’s key building blocks are: open data, labelling of content using semantic standards, and a graphics database that stores and visualizes linked information.

As the third person to speak on the topic, Knud Möller, who is responsible for the open data portal daten.berlin.de, reported on a prototype for linked open data in Berlin. He is testing how GitHub can be used here. The software has been widely adopted and is both open source and free. In addition, no resources are required for setup or hosting. The prototype converts geodata on “Lebensweltlich Orientierten Räumen” (LOR, Life-Oriented Spaces) into linked open data. LOR data were chosen as the input for the prototype since they are the basis for various local-level planning processes.  

Finally, Raimond Spekking from kdvz Rhein-Erft-Rur discussed a survey aimed at developing a standard administrative procedure that local authorities can use to publish open data.

In another brief talk, Dénes Jäger from the Open Knowledge Foundation presented the project Open Data Knowledge Hub, which collects and processes findings relating to open data.

The Municipal Open-Data Network Meeting is a joint initiative of the Bertelsmann Stiftung and the German Institute of Urban Affairs. This digital venue is primarily aimed at stakeholders in local government. At the beginning of the event, participants get to know each other during a speed-networking session. The speakers then spend approximately 15 minutes each sharing new ideas based on their own experience. There is plenty of time afterwards for questions and discussion.