Factorial Ecology in Space and Time of a City

Abstract

The present paper treats the periodic change of the distribution and characteristics of population in a contemporary Japanese City in cycles of 24 hours, i. e. ‘the daily rhythm of a city’. As a typical case, Himeji City is dealt with; the analysis is carried out in a way analogous to factorial ecology.The area which is treated is defined by the municipal boundary of Himeji City. Himeji City consists of an old castle town including a central business district, coastal industrial districts with some centrality, and suburbs which are partly rural in terms of land use. Thus Himeji's multi-nuclei urban structure is obvious in view of journey-to-work patterns (Figure 2).Taylor and Parkes (1975) and others pointed out that traditional studies using factorial ecology had treated exclusively residential (and hence night time) population. Lack of data has been an obstacle to the treatment of population and its change during the day. In this paper, data from a person-trip survey, which was carried out in autumn 1978, are used. From this data, the population was calculated for each time of a day at intervals of an hour, for 51 unit areas of Himeji City, which was classified into subpopulations.The characteristics and the spatial distribution of the population at nighttime (3:00 a. m.) and that of daytime (10:00 a. m.) are shown in terms of factorial ecology. First, nighttime population data from each of 51 unit areas are inputted to factor analysis. Each input variable is a component ratio of subpopulation to the total population of the unit area; subpopulations mean groups sorted by sex, age, industry, and social subgroups. Then, daytime population data, which consist of the same variables set, are inputted. Five factors are extracted, both for nighttime and daytime respectively.The five nighttime factors are interpreted as follows.Factor 1: occupational composition characterized by the contrast between service and factory workers.Factor 2: age composition.Factor 3: industrial composition representing heavy industry workers by negative scores.Factor 4: sex ratio, which suggests also the midnight activity of heavy industry by negative (male) scores.Factor 5: age composition representing old people over 65.Figures 3-5 display the spatial distributions of the first three nighttime factors.The five daytime factors are interpreted as follows.Factor 1: central business activity, particularly highly loaded by service workers.Factor 2: contrast between employed and unemployed people.Factor 3: educational activity.Factor 4: light industrial activity.Factor 5: primary industrial activity by negative scores.Figures 6-8 show the spatial distribution of scores of the first three daytime factors.Since the above-mentioned factor analysis was applied to nighttime data and daytime data separately, we cannot compare the scores between two points of time in a day. In order to compare the population chracteristics between them, three crude variables are selected from the input variables set: these are factory workers, service workers, and housewives, which were the most representative of the variation of the population characteristics among parts of the city at both two points of time in the day.Figures 9 and 10 show the component ratio of these three subpopulations of the 51 unit areas for nighttime and daytime respectively. They suggest that in the daytime, unit areas become more differentiated in terms of population composition than at night.Figures 11-14 show the population for each time of a day at intervals of an hour for four unit areas selected from among 51 unit areas. Each population is classified into seven social subgroups. The Joson area (Figure 11), which is part of the central business district, has a larger population of all the social subgroups in the daytime than at night