Data & Smart City Governance using the example of air quality management

Data governance aims to balance the conflicting interests of stakeholders in the collection and use of data. This may be personal data, data containing commercial secrets, or open data. To enable the use of such data, data governance must ensure the protection of trade secrets, prevent surveillance and censorship, or protect against exploitation, discrimination or political repression. This requires the coordination of all stakeholders at the technological, organisational and normative levels. 

To this end, data governance establishes a set of rules, criteria and indicators, processes, roles, responsibilities, guidelines and standards. These are combined in a data governance concept. 

In our project, the specific data governance concept is developed inductively (bottom-up) on the use case of air quality management in Berlin. At the same time, existing governance principles in the areas of smart city, corporate and data governance will be evaluated deductively (top-down). The result is a data governance concept for data-driven public administration, that can be transferred beyond the specific use case to other municipalities and urban challenges.

There are a number of measures that public authorities can take to improve air quality. In general, the chosen measure should have the greatest possible impact while minimising interference with citizens' civil liberties. To plan such measures, municipalities could rely on data and models. However, these are often provided by private companies. Sharing data or models can be challenging due to privacy issues, trade secrets or commercial interests. What legal frameworks can help resolve the conflicting interests of stakeholders? How can data-driven action and innovation be enabled in a way that serves the public good?

A municipal measure should take into account the interests of all stakeholders. This includes the needs of the affected citizens, associations or other actors in urban society as well as economic actors and political interests. Therefore, the data or models used to make decisions should not be accessible only to administrations or private companies. Citizens need access to key information in order to participate in (democratic) decision-making processes. A data governance concept must therefore also determine when and to what extent public participation should take place and what added value it can offer both the administration and the citizens themselves. In addition, the organisational feasibility of the planned measure must be assessed. It is important to ask in advance whether a public administration can reasonably implement it within the given timeframe. What internal decision-making procedures are in place for the implementation of the measure, and how can these be adapted or supplemented if necessary?

At an organisational level, we also look at the impact on companies that work with governments. For example, in terms of data quality requirements, the explainability of algorithms, and open vs. proprietary source and data strategies.

Finally, it is necessary to assess whether the measure taken by the public administration is technically feasible in the light of the normative and organisational requirements. It should be considered  whether the implementation is proportionate and reasonable. In addition, it is necessary to determine whether the measure can actually achieve the desired objective (e.g. reduction of air pollution) and which technological adjustments can improve the achievement of the objective.

In order to implement data-driven measures as effectively and conflict-free as possible in the future, the project examines which existing legal regulations and organisational procedures can and should be adapted or supplemented. Our focus is on identifying changes that can be implemented easily and effortlessly ("low hanging fruit first").