Multidimensional analysis of social discourse on Twitter (X) around Venezuelan migration in Colombia

Colombia has become the principal host of Venezuelan migrants in Latin America, prompting a complex mix of solidarity, anxiety, and discrimination. This study offers a multidimensional analysis of Twitter (now X) discourse on Venezuelan migration in Colombia between 2014 and 2022, using natural language processing, geolocation inference, sentiment analysis, and manual annotation of derogatory terms. Drawing from 7.4 million posts, we examine how online discourse evolves across time, space, and language. While negative sentiment is pronounced in original tweets during periods of political and social unrest, our findings highlight that users tend to amplify neutral and positive messages-over 67% of retweets fall into these categories-suggesting an informal counterbalance to polarized discourse. Additionally, the analysis of the derogatory terms veneco and veneca reveals significant gendered patterns: the feminine form veneca appears more frequently in negative contexts, pointing to the gendered dimensions of xenophobic language. These results underscore the dual function of social media as a site of both stigmatization and resistance, offering insights into the emotional and cultural dimensions of migration discourse in Colombia's digital public sphere.