Social media analytics (SMA), referring to the collection and analysis of user generated data from social media platforms, attract attention of both researchers and practitioners striving to derive consumer insights. The SMA domain grows multifariously, with a highlight on the capability of machine learning algorithms in capturing noteworthy insights through processing high-volume and complex data in a cost effective way. As machine learning applications draw attention as a fertile area that may re-shape the future of SMA, there is a need to comprehend trends and approaches in an integrative framework. Accordingly, this study aims to present an integrative framework by portraying machine learning application trends and approaches in SMA. 42 scientific articles published in refereed scientific business, management, and computational science journals between the years 2013 and 2019 are analyzed via systematic literature review based on visual text mining method (SLR-VTM). The results revealed five distinctive research clusters as: (1) review sites, (2) microblogs, (3) social networking sites, (4) content communities, (5) cross-media. This analysis plays a crucial role for enhancing our understanding regarding the intellectual structure of the field, acknowledging the leading studies of the domain, better positioning future research, and determining gaps and new paths for researchers.