The pitfalls of city data strategies

By Bas Boorsma, Vice President EMEA, Cities Today Institute*

Cities have adopted and implanted evolving urban digitalisation strategies, and most of those increasingly focus on data. In fact, data is at the heart of the latest chapter in smart city innovations, following the years of a focus broadband infrastructure, vertical point solutions and Internet of Things pilots. Data strategies encompass many components tied to data: algorithms, open data approaches, artificial intelligence, and data governance among others. As many cities have experienced, this latest journey is full of pitfalls and challenges.

The challenges faced in articulating and implanting data strategies for cities mirror the issues faced in the smart city space at large: rich on hype, lack of clarity on definitions, assumptions that may be false but have a habit of living on for a very long time, and experiments without a defined purpose.

First, data strategies all too often serve generic innovation goals that leaders hope to meet by means of a generic convergence of data. However, without a clear purpose articulated, such initiatives may falter, or under-perform at best. Many smart city open data initiatives suffer from this ailment: it is often assumed that if a lot of data is gathered this will somehow be a force for good, with a hackathon on top of it producing magical innovations that we can only hope will be relevant. Open data initiatives that are defined with a clear purpose in mind (such as next-generation multimodal mobility) have a much larger chance of success.

Second, all too often it is assumed that ‘data is the new oil,’ with data representing an asset that carries value. The reality is that it often does not. There is no scarcity of data, and a simple sample of data often suffices to achieve what needs to be achieved. Any innovation or business model that is based on the assumption that data will get monetised deserves healthy scrutiny: sometimes it can; often it cannot or cannot yet.

Third, data–more than any software or hardware–can be found at the heart of ethical considerations that come with digitalisation. Without an agreed set of rules and parameters as to whose data gets used under what conditions and how it will be managed post prime or secondary use, smart city project leaders find themselves in the middle of controversy and debate.

Fourth, building on the previous point, data governance often gets fuzzy as smart community data strategy leaders converge different and often conflicting goals in their data governance structures. Many open data initiatives serve the purpose of innovation yet also become constructs that provide oversight on ethical use and regulations. I maintain that these do not go well together. Ideally such functions are clearly split: for example a data utility or a data commons can be set up to provide rules, regulations and ethical oversight on the one hand, and on the other a data for innovation exchange to drive innovation–different missions, different organisations, different professionals at work.

Fifth, attempts to provide oversight on data use within community digitalisation strategies have all too often been too limited in scope: the lesson learned is that any comprehensive attempt to regulate should not be restricted to data or data use itself, but should include algorithms and platforms also. Further, such oversight should be introduced in the earliest of pilots and proof of concepts–the same goes for cyber security–as early proof of concepts have a nasty habit of living on for a very long time.

Sixth: the assumption exists that data somehow magically translates into actionable insights. Generally it does not: analytical models are needed, purpose must to be defined, context needs to be understood, and deep learning helps us peer into what otherwise strikes us as insurmountable complexities. All the above elements are the stepping stones on the road towards insights. And even if insights are obtained, they not equate action. Beyond obtaining insights we need to decide how to make them actionable and turn them into policy. The equation that reads like this: “data, arrow, insight plus policy” may work in marketing departments, but the reality on the ground is substantially more complex.

Seventh, much of the data you need may already be under your nose. Try to obtain clarity as to what data you need exactly to solve the challenge at hand or that is needed to fuel a solution you seek. Next, many discover the data required may already be available and within reach, rather than requiring exotic public-private exchange arrangements.

Last: like data itself, there is no scarcity of hype and marketing lingo. For years we have been tormented by the much abused term ‘big data’ (whereas most of us just meant data or actually needed ‘small data’). ‘Data powers cities.’ ‘Data is the new oil.’ ‘Open data.’ ‘We don’t see sentiment, we see behaviour.’ I think we all recognise the litany of terms and phrases that may (or may not) touch on valid points but remain without sufficient validation, or definitions collectively agreed and understood.

*This article was derived from the Revised Edition of A New Digital Deal. Beyond Smart Cities. How to Best Leverage Digitalization for the Benefit of our Communities (January 2020)