Uncovering Trends in Human Trafficking and Migrant Smuggling Activities: A Natural Language Processing Approach to UNODC SHERLOC Database (original) (raw)

Human trafficking affects every country of the world and it often occurs from less to more developed countries, where people are rendered vulnerable to trafficking by virtue of poverty, conflict or other conditions. A similar trend in human mobility is identified in migrant smuggling activities, which by definition differs in that it involves the procurement, in order to obtain a financial or other material benefit, of the illegal entry of a person into a state without however implementing the component of exploitation. However, when cases are manifested, these two issues have often been addressed through interchangeable legal frameworks depending on the country, the individual case and the political context. In this project we bring together 2,284 collected case summaries of judicial decisions categorised either as human trafficking (n = 1490) or migrant smuggling cases (n = 755) or both from 133 different countries, as indexed in the UNODC SHERLOC database, in order to explore similarities and differences in the prevalent criminal activities while also focusing on the countries' geographical position and demography. The goal is to identify whether there is (or not) a trend in the manifestation of the involved criminal activities. The research pioneers the incorporation of methodological steps which include Natural Language Processing (NLP) techniques to identify topics in the mined text (topic modelling) on the UNODC SHER-LOC's Case Law Database. We further identify and codify topics on the provided Fact Summaries (descriptions of the legal cases) in order to examine prevalence of criminal activities for each case and across countries.