Extraction Urban Clusters from Geospatial Data: A Case Study from Switzerland

Abstract

Different techniques were developed to extract urban agglomerations from a big dataset. The urban agglomerations are used to understand the structure and growth of cities. However, the major challenge is to extract urban agglomerations from big data, which can reflect human activities. Community urban cluster refers to spatially clustered geographic events, such as human settlements or activities. It provides a powerful and innovative insight to analyze the structure and growth of the real city. In order to understand the shape and growth of urban agglomerations in Switzerland from spatial and temporal aspects, this work identifies urban clusters from nighttime light data and street network data. Nighttime light data record lights emitted from human settlements at night on the earth's surface. This work uses DMSP-OLS Nighttime light data to extract urban clusters from 1992 to 2013. The street is one of the most important factors to reflect human activities. Hence, urban clusters are also extracted from street network data to understand the structure of cities. Both of these data have a heavy-tailed distribution, which includes power laws as well as lognormal and exponential distributions. The head/tail breaks is a classification method to find the hierarchy of data with a heavy-tailed distribution. This work uses head/tail breaks classification to extract urban clusters of Switzerland. At last, the power law distribution of all the urban clusters was detected at the country level.