Topic Modeling and Characterization of Hate Speech against Immigrants on Twitter around the Emergence of a Far-Right Party in Spain

Abstract

In this paper, we sought to model and characterize hate speech against immigrants on Twitter in Spain around the appearance of the far-right party Vox. More than 240,000 tweets that included the term ‘Vox’ between November 2018 and April 2019 were automatically collected and analyzed. Only 1% of the sample included hate speech expressions. Within this subsample of 1977 messages, we found offenses (56%), incitements to hate (42%), and violent speech (2%). The most frequent terms used were classified into five categories: Spain, Immigration, Government, Islam, and Insults. The most common features were foul language, false or doubtful information, irony, distasteful expressions, humiliation or contempt, physical or psychological threats, and incitement to violence. Using unsupervised topic modeling, we found that the four underlying topics (control of illegal immigration, economic assistance for immigrants, consequences of illegal immigration, and Spain as an arrival point for African immigrants and Islamist terrorism) were similar to those in the discourse of Vox. We conclude that the hate speech against immigrants produced around Vox, and not necessarily by Vox, followed the general patterns of this type of speech detected in previous works, including Islamophobia, offensive language more often than violent language, and the refusal to offer public assistance to these collectives.

Refugees Welcome? Online Hate Speech and Sentiments in Twitter in Spain during the Reception of the Boat Aquarius

“…This shows the relevant role played by stereotypes and prejudices, if not complete lies or fake news, such as the connection of immigration with terrorism. This is coherent with the observations made in previous studies [15,17,20], and highlights the need for initiatives and studies that approach the interaction between hate speech and disinformation [31,44]. In general, the topics associated to hate speech around the studied event are strongly connected with the frames observed by Lawlor and Tolley [57] in the media representation of immigration.…”

“…This way, the study offers preliminary observations that need further analysis and confirmation from future studies, as well as an update of the predominant sentiments and topics, given that, despite the milestone condition of the arrival of the Aquarius boat, the conversation around the topic of asylum and migration might have changed significantly, especially after the appearance of a right-wing anti-immigration party such as Vox [15].…”

“…This will allow us to fill in a still existing knowledge gap on the feelings of the Spanish population towards refugees and immigrants, not only around this specific high-profile event, but also in the context immediately before the arrival of Vox to Spanish institutions in December 2018, which made anti-immigration a more relevant topic in the Spanish political discourse [18]. This way, the study will complement previous works in the field and try to observe how and around what topics the anti-immigration discourse in Spain is built [15,18,19] and, more in general, how refugees are depicted in Twitter [20,21]. Moreover, the article introduces a novelty in its approach to hate speech against politicians, even if they are not a vulnerable group that demands protection, one of the most common reasons behind hateful contents is political ideology.…”

“…With this in mind, the study will use text mining, or text analytics, not only for the need of overcoming the limitations of surveys in racism-or xenophobia-related issues [11], but also since the access to unstructured data, such as the one of Twitter, provides unfiltered and immediate information in a particular moment in a way that surveys or other studies cannot. Text mining refers to a set of techniques that are gaining relevance in different fields of the academia, from finance [12] to social impact [13], including politics [14] and antiimmigration sentiments [15], the topics of this study. From these techniques, which offer a way to extract knowledge from textual information using big data [16], sentiment analysis and topic modelling, the two that will be used in this study, are among the most common ones, both with a growing presence, according to the systematic review of literature on text mining on Twitter conducted by Karami et al [17].…”

Using Twitter to Track Immigration Sentiment During Early Stages of the COVID-19 Pandemic

“…Hate speech detection seeks to identify and classify threatening, harassing or seriously offensive language. Several studies have employed Twitter data to study hate speech against immigrants (Bosco et al, 2017;Sanguinetti et al, 2018;Basile et al, 2019;Comandini and Patti, 2019;Calderón et al, 2020). Hate speech is seen as a specific expression of sentiment or attitudes and thus related studies have focused on characterising how the social network expresses hate.…”