357 / 2018-12-16 18:30:43

Explore the Relationship between Diversity and Density of Urban Activities by POI-based Measurements: Case Studies of Three Typical Chinese Cities

Urban activities, diversity, density, POI

全体主题 > Spatial statistics

Although there is an extensive body of literature that demonstrates the density of urban activities is empirically correlated to the diversity, such relationship is seldom quantitatively shaped by scholars because there is no consensus on the measurements of diversity. As the most commonly adopted method, Shannon Entropy, has been gradually recognized having some limitations in gauging the diversity, in particular for measuring the multidimensional features of diversity. In addition, the inadequate and to some extent inappropriate data (For example, Land use data) is another obstacle for exploring the relationship. Nevertheless, newly emerging data sources such as Point of Interest (POI) provide another possibility for developing new indices of diversity and density of urban activities to interpret their correlation. Taking advantage of POI data extracted from online navigation map, the dissertation utilizes the counts of POI in unified grids as a proxy of the activities’ density and used Hill Numbers, a framework containing a series of diversity measurements, to display the multidimensional diversity of urban activities. Here, Hill Numbers is originally derived from the ecology, which includes Richness, Shannon, Simpson and other indexes for describing the degree of diversity. Adopting such explicit data sources and the Hill Numbers provides the author a better perspective for exploring the spatial correlation between the density and diversity. Primary explorations on the Shannon and density maps and three GWR (Geographic Weighted Regression) models for different scale cities are presented. The results initially reveal that the density can be to some extent explained by the diversity measured by Hill Numbers as the global R squares in three cities are all above 0.65. Moreover, the results also show the inadequacy of traditional Shannon Entropy for measuring the diversity of urban activities. Furthermore, some detailed but interesting findings such as the comparisons between the locations with high Shannon index but low density and high density but low Shannon index are also explained in the dissertation. These to some extent insightful findings might be useful for the guidance of urban planning and community design.