Skip to main content

Using Big Data to study violence against women and girls and its challenges online (Italy)

Languages and translations
File type1
E5_WP19_Villante_EN.pdf (application/pdf, 454.82 KB)
File type1
E5_WP19_Villante_RU.pdf (application/pdf, 485.47 KB)

Social networks are the "modern agoras” and, during the pandemic, it has become a social space where violence against women increased by moving from real to virtual. Social media are the new source of data for understanding, along with other data sources, the changing nature of gender-based violence. Finding methodologies and techniques able to use these new sources of Big Data is essential to understanding and monitoring that phenomenon. Observing what is happening on social media, and measuring the sentiment of online conversations, however, is not the only goal. Indeed, we know that it is violence against women itself that has become virtual (online violence) and finds its channel of dissemination in social media. ISTAT has piloted an Index related to gender-based violence and a sentiment and emotion analysis based on social media content (Twitter, Facebook, Instagram, and web newspaper post) aimed at measuring how the users of social react to the violence against women and/or generates discussions around it. In addition, the adopted method helps to understand which are the messages that unleash the discussion, by providing a map of the topics (such as body-shaming, femicide, and rape) where the sentiment is most solicited (both positive and negative). The results of the experimentation are furthermore compared to another source of data, coming from the national helpline against violence and stalking against women (1522) that ISTAT has recently adopted (from 2020) as a new and timely administrative source of data in order to gather information during the pandemic period. The piloted method can help researchers in monitoring the future challenges of violence against women and girls online and opening a new way aimed at integrating different sources of data, as our study has attempted to provide.