This talk overviews ongoing work on information diffusion in social media, focusing in particular on mining, modeling, and prediction tasks on data from the Twitter network. I present machine learning efforts that leverage the structure of meme diffusion networks and many other features to detect misinformation campaigns, such as astroturf and social bots. We use agent-based and maximum-likelihood models to understand the formation of communities, the creation of social ties, and the competition for attention. We investigate how different forms of structural and topical diversity in the network can be leveraged to predict which memes will go viral. Finally, it time permits, I will review a few crowdsourcing projects exploring the computation of unbiased scholarly impact metrics, the anonymous collection of sensitive data, and social good. Joint work with many members of the Center for Complex Networks and Systems Research at Indiana University (cnets.indiana.edu). This research is supported by the National Science Foundation, McDonnell Foundation, and DARPA. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of these funding agencies.