Overview

Data science in local government involves using novel techniques, such as machine learning, to harness and analyse both internal administrative data and especially data from third parties to make government more responsive to citizens and more efficient in targeting resources. For example, Lyon is one of almost 100 cities using data from the Strava cycling app to plan cycle routes; whilst Cardiff is making use of Vodafone telephone data to help manage traffic congestion, and London is using Oyster card data to optimise its underground system. The spread of data science promises to revolutionise the ability of local policymakers (who have typically operated in an information poor environment) to plan and manage their communities. It is potentially particularly revolutionary for small and medium sized cities, as it offers an alternative way of realising the “smart city” vision which does not rely on the installation of expensive sensor grids.

However, the novelty of data science as an area means that there has been almost a complete lack of academic attention to the subject. We know little about why some cities choose to take up data science methods whilst others do not. We also know little about how to overcome barriers to implementation (for example, training key staff, privacy and data protection, and contracting IT companies). Finally, there has been a lack of critical attention to the impact of novel algorithms and machine learning techniques in local government, especially when they are based on data which comes from third parties. Does data science benefit some citizens more than others? How does it alter power relations within government, and between government and the private sector? The aim of this project is to help fill in these gaps in knowledge, by both explaining the spread of data science in the local government context and understanding its impact.

Project website