Abstract
Abstract
Network characterization of goods movement is particularly important to estimate needs and accurate management for freight transportation. The present study aims at characterizing the network of goods movement in Indonesia and examining the effectiveness of network characterization for prediction purpose. A network approach using RStudio was applied. It was found that the network of goods movement in Indonesia is characterized by weak inter-provincial goods movement, scale-free network, disassortative, having rich-club phenomenon, and having a core-periphery structure in which Sumatera Barat (West Sumatera) and Jawa Barat (West Java) appears to be the cores of the network. The findings also demonstrate that the prediction model developed by characterization performs better and can explain, on average, 39% of the variances. The prediction models for chemicals, fuel, meat and livestock, fish, crude oil, and fertilizer, have even better capability to explain variances in the volumes of goods movement by more than 60%. Potential avenues for future research are also discussed.