Network Visualization and Anomaly Detection in International Timber Trade Flows

Abstract

Our planet’s forests, which are vital for sustainable development, are threatened by deforestation around the world. The problem of missing timber flows poses a significant barrier to preserving our remaining forest reserves, maintaining biodiversity in these ecosystems, and promoting sustainable and fair trade of timber. In this thesis, we utilize (hitherto underutilized) international trade data and historical rates of tree cover loss in major tropical countries to compare trends across datasets in the global timber supply chain and identify anomalies in trade patterns that can be attributed to technical and non-technical errors (that are often indistinguishable).

We first focus on the primary trade partners of our case study countries, Brazil and Indonesia, across three timber trade datasets. We then analyze discrepancies between reported exports and imports for trade partners and compare them with illegality estimates for tropical timber trade. We find a high correlation for the early 2000s between the discrepancies in the FAOSTAT Forestry Trade Flows timber trade dataset and illegality estimates from Chatham House. We further evaluate the similarities over time between timber trade discrepancies and illegality estimates using a map-based network visualization prototype we designed and built. Our work suggests the need to synthesize accurate trade data on global timber flows for regulating illegal activities at source sites and in the supply chain of timber products.