Designing and developing an application for online hate speech detection on social media

Abstract: Social media platforms have a significant impact on the world. Through connecting people over the globe, social media has resulted in multiple positive effects. However, the impact is not only beneficial. Social media platforms have also caused polarization and criminal uprisings, leading to events such as the Rohingya genocide in Burma. By tracking the discourse on social media channels, it is possible to get insights into societal movements and warnings about what is to come. However, the massive number of conversations that occur on social media channels can be hard to analyze. Especially for NGOs, which commonly has strained resources and budgets. This study aims to research how tools can be developed to help NGOs and similar organizations in their work towards finding threats from online activity, as well as analyzing the general discourse in a subject. The contribution of this research is twofold: (1) The paper investigates how the NGO Civil Rights Defenders is currently working with analytics of social media and how their work can be helped. This is done by interviews with Civil Rights Defenders, and the results are presented in two user scenarios and use cases. (2) A tool is developed for helping Civil Rights Defenders improve their work within social media analysis. This is done through a literature review of adjacent research and the current state of technology, as well as using insights from interviews. The collected information is then used to build a fully working web application, targeted mainly towards NGOs and their work in fighting hate speech online. The developed tool was tested, and the results were satisfactory. The testers were able to use the tool without much, if any, confusion, and the tool provided help in their workflow.