A city data brokerage is the next step for open data and provides a means for individuals and organisations to publish, buy, sell and trade data. It provides an essential level of trust and control to encourage organizations to make available high value data sets. It also offers a framework in which individuals can begin to share and control access to their personal data.

One of the forefront cities in exploring this concept is Copenhagen, who have deployed a trial data brokerage since 2016. We blogged about the project in 2016 in http://urbanopus.net/smart-city-copenhagen-key-lessons-and-future-directions/

Recently, Copenhagen have published some of their initial finding and lessons learnt in a report that makes interesting reading for those trying to understand how to go beyond the open data model.

Some key facts are:

  • Data broker built and managed by Hitachi as a PPI type initiative
  • 140 different datasets available for purchase/use
  • Over 1000 individuals/organizations involved in the trial
  • Significant outreach (workshops, hackathons, 1 on 1s) to understand the data landscape

Some key learnings:

  • data brokerage is still an immature market and needs well developed use cases

Interestingly, the main stumbling block for data sharing is a lack of clear uses cases or exemplars for potential participants. Owners of data are aware that the data has value, and are willing to consider sharing, but they want some clear use cases to help them identify potential buys and help them put a value on the data. Equally, potential data consumers are looking for clear use cases before they commit to buying/using data. A clear lesson, which we already know from the open data movement, is that publishing data in the hope that people will use it, doesn’t really work. Copenhagen heard a clear message that use cases differ significantly for any data set and data needs adapting, augmenting, cleaning in different ways for different users. They offer an interesting example of people movement patterns. While many potential users agree they would like access to such data, what each means by ‘people mobility data’ is very different, eg real time versus historic, granularity of data, geographic spread, data format etc all differ depending on the use case.

  • Create a data competence hub

This lesson grew out of their experiences working with data providers and consumers. While there was a significant interest in data sharing and brokerage, a significant amount of discussion was needed to engage and work to partner organizations. This was partly because of the immaturity of the market, but also because of trust and liability issues. As such, the project spent considerable time working with potential data users, and matchmaking via individual or group meetings. In a similar fashion, and related to the use case point above, it was clear that data often needed to be adapted or combined before  it became useful to potential users. Again this required discussion around tools and techniques to work with data.

A significant outcome of the project was the realization that a technical data sharing platform, ie the data broker, was a necessary requirement for data sharing, but not sufficient on its own. It needs to augmented by a mechanism or approach that allowed data providers and consumers to engage and negotiate data formats, details and usage. The project has highlighted the notion of data collaboratives – marketplaces where data providers can interact with potential buyers and discuss how to collaborate around data sets. (see http://datacollaboratives.org/)

  • Create simple guidelines and standards for data publishing

A final lesson was the need to develop guidelines for data publishing, and agree on common formats. This is driven out of the difficulty many potential data users had in accessing and using data. This was partly due to formatting issues with available data – the ubiquitous use of PDF in open-data sets is a common issue – since PDF is a difficult data format to access, tease apart and reuse. However it was also related to the need for toolchains around data that make it simple to access and use the data – for example, a common request was basic visualizations of data to allow potential users to explore data before buying.


The lessons from 2 years of real data brokerage in Copenhagen are important for any City looking to move beyond the simple provision of open data. Innovation and improved services are possible through better use of data, but it’s not as simple as making the data available and assuming great things will happen. A careful program of technology platform (the data broker) and support via use cases and date community development is also needed.