The increasing availability of digital datasets via the web and social media have provided social scientists with unprecedented opportunities to study individual and group activity at scale. New (sub)disciplines such as digital humanities, computational social science, and digital sociology have emerged out of this landscape, providing an exciting research environment where the social meets the computational. In this seminar I highlight and discuss some of the key opportunities, challenges, and open problems for using data science methods in social science research. To explore this space, I draw on examples and experiences from my own research collaborations, in particular a recently published study that combines advanced computational methods with social theory and traditional qualitative analysis to map and study the anti-vaccination movement on Facebook. I focus on three methods in particular: social network analysis, topic modelling, and gender modelling using natural language processing and historical census data. In doing so, I seek to draw attention to how computational social science involves much more than simply throwing data science techniques at social data.