빅데이타 분석을 통한 자동차산업에서의 클러스터 형성과 생태계 연구

This study tries to make company clusters in the automobile industry using more than 40 thousands transaction data provided by Korean Enterprise Data for the year of 2011, to find out the characteristics for the industry ecosystem. We find out 45 company clusters based on the transaction closeness between companies using the tools of social network analysis. Results show that there exist one giant cluster occupying 31 sales share of all firms in the network. But we find out that there exist middle size clusters consisting of one automobile assembly company and cooperative material and parts providing companies and also clusters consisting of single kind of parts/material companies. Results also imply that most of clusters have the hierarchical tree-like transaction patterns centered on single or two firms playing central role in each clusters. Finally we checked if these clusters have sales independence from big companies, especially from the hub company of the giant cluster using sales dependency. We find that the hub company of the giant cluster have overwhelming influences on other clusters. But we also find there exist the independent clusters from this company. We suggest that the industry policy focus on the effectiveness of the transactions for the independent clusters and that the policy focus on diversifying the transactions leading to the fairness of the transactions for the clusters heavily dependent on a few companies.