Mapping social conflicts in natural resources. A text-mining study in mining activities

Abstract

Applying text-mining techniques we have developed a methodology that measures the number of social conflicts related to the exploitation of nonrenewable natural resources. We focused on conflicts in four mining countries (Australia, Canada, Chile, and Peru) from 2003 to 2016 considering more than 20,000 articles from the major newspaper of each country. From our data we detected cross-country and crossregional differences and changes in time patterns. We found a statically significant correlation between our main index and mineral rents in % in GDP. However, our results should be interpreted with caution since we have not taken into account endogeneity issues and our indexes could be biased by different level of lobby powers among our country sample. Our main contribution is the generation of novel database with different indexes of soft conflicts related to the exploitation of non-renewables natural resources and its media coverage in Australia, Canada, Chile and Peru.