Substitutability of urban sustainability assessment indicators: A semantic similarity analysis

Abstract

Urban sustainability assessment is an effective method for objectively presenting the current state of sustainable urban development and diagnosing sustainability-related issues. As the global community intensifies its efforts to implement the sustainable development goals (SDGs), the demand for assessing progress in urban sustainable development has increased. This has led to the emergence of numerous indicator systems with varying scales and themes published by different entities. Cities participating in these evaluations often encounter difficulties in matching indicators or the absence of certain indicators. In this context, urban decision-makers and planners urgently need to identify substitute indicators that can express the semantic meaning of the original indicators and consider the availability of indicators for participating cities. Hence, this study explores the relationships of substitution between indicators and constructs a collection of substitute indicators to serve as a reference for sustainable urban development assessment. Specifically, building on a review of international and Chinese indicators related to urban sustainability assessment, this study employs natural semantic analysis methods based on the Word2Vec model and cosine similarity algorithm to calculate the similarity between indicators related to sustainable urban development. The results show that the Skip-gram algorithm with a word vector dimensionality of 600 has the best performance in terms of calculating the similarity between sustainable urban development assessment indicators. The findings provide valuable insights into selecting substitute indicators for future sustainable urban development assessment, particularly in China.