• January 15, 2019
• Civic Analytics Network

The 2010s have seen a rapid expansion of executive-level data and analytics positions in local government. Following the rise of the “chief data officer” in both the private and public sectors, and what many see as inadequate urban policy leadership at the federal level, expectations run high for the impact data analytics could make in cities. To help fulfill this promise, a patchwork of universities, philanthropies, and non-profits are supporting cities in adapting analytics innovations developed in one city to others across the country. 

How does an algorithm that prioritizes smoke alarm distribution in New Orleans change the way a similar service is delivered in Syracuse? Why does a statistical model for rodent mitigation developed in Chicago fail to work in Pittsburgh? How should a data-processing solution in Louisville, or a problem-solving methodology in New York City, be iterated in other cities? This paper—developed from academic literature review, interviews with analytics practitioners in local governments, and notes from roundtables of experts—explores both the theoretical solvency and practical considerations for replicating data-analytics use cases from city to city. 

The findings are mixed. Stories of analytics replication are often told in terms of their potential rather than realized value. Both “analytics” and “replication” suffer from loose definitions, and applications of data science in cities—even in the largest cities with the most sophisticated analytics programs—are often proofs of concept, rather than a core component of performance management and innovation delivery. What is clear is that data analytics solutions are far from a silver bullet to public problems, and there is substantial work that must be done to develop the way analytics “successes” are quantified before the field can have a productive conversation on inter-city analytics replication at scale.

The first section of this paper contextualizes the topic of civic analytics replication in the fields of technology transfer and urban policy diffusion, arguing that "adaptability" is a more appropriate conceptual framework for reconfiguring analytics use cases in new contexts. A template of factors to consider when attempting to adapt a project from one city to another follows. Brief case studies illustrate the potential and challenge of adapting analytics projects, and the paper concludes with suggestions on types of projects to adapt and recommendations for the field.