A hierarchical regionalization of Japanese prefectures using 1972 interprefectural migration flows

Abstract

Slater P. B. (1976) A hierarchical regionalization of Japanese prefectures using 1972 interprefectural migration flows, Reg. Studies 10, 123–132. A methodology for utilizing an interregional transaction flow table to construct a hierarchical system of regions is presented. The square matrix is first adjusted by a biproportional procedure to possess uniform row and column sums. A series of increasingly sparse digraphs can then be associated with the table by raising threshold levels which are used to convert it into (0, 1) matrices. The strong components of these digraphs (Boolean adjacency matrices) can be hierarchically arranged. 1972 Japanese interprefectural migration data are examined in this manner. Several officially recognized regions (chiho)—Shikoku, Chugoku, Tohoku, and Kyushu—are readily distinguished, since they correspond to strong components which are stable over a wide range of thresholds. The chiho of Kanto, Chubu and Kinki are not identified, however. The Seto-Uchi area, San-in and several apparently previously unrecognized collections of prefectures emerge as well-defined regions which have relatively large internal interprefectural migration.

Keywords:

• Hierarchy of regions
• Clustering
• Graph theory
• Strong components
• Biproportional adjustment
• Flow tables
• Migration
• Japanese prefectures

Additional information

Notes on contributors

Paul B. Slater

Daniel R. Vining, who provided the data analyzed, also gave many comments on an earlier version of the paper. The National Science Foundation under Grant No. SOC74-23880, ‘Applications of Graph-Theoretic Clustering and Transaction Flow Analysis in Geography and Regional Science,’ furnished financial support for this study.