10.7910/DVN/IYOIDDTong, XinXinTongDuke Kunshan UniversityLi, JiayiJiayiLiDuke Kunshan UniversityLi, YixuanYixuanLiDuke Kunshan UniversityZhang, LuyaoLuyaoZhang0000-0002-1183-2254Duke Kunshan UniversityHarvard Dataverse2022Arts and HumanitiesComputer and Information ScienceSocial SciencesZhang, LuyaoDuke Kunshan University2022-01-162022-01-1629316014022873150text/csvtext/tab-separated-values1.0CC0 WaiverMinority groups have been using social media platforms to organize their social movements that create profound social impacts. Black Lives Matter (BLM) is one of the many successful social movements that started and expanded on social media. However, quantitative analysis with rigorous machine learning and natural language processing is absent. We apply Latent Dirichlet Allocation (LDA) model to analyze more than one million tweets with #blacklivesmatter following a major BLM movement and compared the results to those with #stopasianhate after a Stop Asian Hate (SAH) event. Our findings revealed that the tweets presented a thorough examination of the most influential topics, including the dominant topic of politics and social justice. By comparing the analysis results from the two most recent and critical online social movements, our study contributes to the topic analysis of social movements on microblogging platforms in particular and social media in general.