Video Demographic Analytics of Social Media Users: Machine-Learning Approach

Nowadays, several social media platforms enable people to upload and share videos of their daily activities and debates. The volume of online video content is increasing at a rapid pace. These videos are very compelling and can convey significant demographic information for social sensing that can help governmental and non-governmental organizations and businesses for many purposes such as strategic planning, political and social development decisions, and market research. Demographic analysis refers to studying the composition characteristics of a group of people. These characteristics may include gender, age, race (ethnicity), dialect (accent), education level, disability, household income, nationality, sentiment, and others. The study of these characteristics can guide informative decisions and have huge socioeconomic impacts on people's life. This chapter reviews work related to multimodal video demographic analytics. Moreover, it presents a multimodal demographic detection system from videos on social media with focus on gender, age-group, and dialect of Arabic speakers as a case study. Several features are investigated for each modality and fused using a machine-learning approach.