Price characterization of import goods from Customs dataset

Abstract

Investigating import activities is advantageous for a country as it provides explanation on various economic phenomena such as product price sensitivity in the local and national scale. In this work, we use the 2016 regulated import lists from the Bureau of Customs to examine price distributions of imported products. By k-means clustering, products were clustered into groups using their percentile z-scores. Principal Component Analysis was used to further separate the distributions based on their heterogeneity. It was determined that the first principal component PC1 describes the presence of outliers, while PC2 corresponds to the variability of the import prices for a specific product.