Mapping local government priorities: a web-mining approach for regional research

Abstract

The relevance of institutions for regional development has been well established in economic geography. In this context, local and regional governments play a central role, particularly through place-based and place-sensitive strategies. However, systematic and scalable insights into their priorities and strategies remain limited due to data availability. This paper develops a methodological approach for the comprehensive measurement and analysis of local governance activities using web mining, natural language processing (NLP), and machine learning techniques. We construct a novel dataset by web scraping and extracting cleaned text data from German county and municipality websites, which provides detailed information on local government functions, services, and regulations. Our county-level topic modelling approach identifies 205 topics, from which we select 30 prominent topics to demonstrate the variety of topics found on county websites. An in-depth analysis of the three exemplary topics, Urban Development and Planning, Climate Protection Initiatives, and Business Development and Support, reveals how strategic priorities vary across space and how counties differ in their framing of similar topics. This study offers an explanatory framework for analysing the discursive dimensions of local governance and mapping regional differences in policy focus. In doing so, it expands the methodological toolkit of regional research and opens new avenues in understanding local governance through web data. We make an aggregated version of the data set freely available online.