Toxic online discourses have the potential to cause disputes or damage communities. Without a formal study of hate speech datasets, existing reviews frequently concentrate on certain hate speech categories. This study highlights datasets, characteristics, and machine learning models while conducting a methodical analysis of textual hate speech detection systems. 138 publications were analyzed, and the findings vary by category. The most common methods combine deep learning models. There is a demand for improved resources since many datasets are limited and unreliable for a variety of activities. This study offers insights into the characteristics of hate speech and makes recommendations for future research paths.https://www.mdpi.com/2078-2489/13/6/273Share this:FacebookXLike this:Like Loading... Post navigation A Systematic Review on How to Address Hatred in its Various Manifestations: Understand Its Different Aspects, Use Different Tools and Specific Interventions (Global Psychiatry Archives) Two Weeks in Soft Security: Free Resources on Countering Extremism, Hate, and Disinformation, March 2025 (II/II)