Detecting industry clusters from the bottom up based on co-location patterns mining: A case study in Dongguan, China

Abstract

Industry cluster development is important to stimulate regional economy. Conventional spatial methods for detecting industry clusters use a pairwise manner to infer the co-location relationships of multiple industrial types or instances, which increases the difficulty of interpreting the results. This study proposes to use co-location patterns (CPs) mining method to directly capture the co-location of multiple industrial types from the bottom up without any conditions of data relations defined a priori. The method is applied in Dongguan, China, to investigate the industry cluster patterns at an intra-urban scale. At the city level, the results show prevalent CPs of information communication and technology industry and other associated sectors. At the sub-regional level, however, approximately 41% of the industrial CPs are different from those obtained at the city level. The local features of sub-regional industry clusters are associated with productions of, for instance, sporting goods and toys, digital instrument and office equipment, machine parts and woodware, and textile-related products.