Social media is becoming an increasingly important data source for learning about and tracking breaking news. This is possible thanks to mobile devices connected to the Internet, which allow anyone to post updates from anywhere, leading in turn to a growing presence of citizen journalism. Consequently, social media has become a go-to resource for journalists during newsgathering. Use of social media for newsgathering is however challenging, and suitable tools are needed in order to facilitate access to useful information for reporting. In this paper, we provide an overview of research in data mining and natural language processing for mining social media for newsgathering. We discuss seven different tasks that researchers have worked on to mitigate the challenges inherent to social media newsgathering: event detection, summarisation, news recommenders, content verification, finding information sources, development of newsgathering dashboards and other tasks. We outline the progress made so far in the field, summarise the current challenges as well as discuss future directions in the use of computational journalism to assist with social media newsgathering. This survey paper is relevant to computer scientists researching news in social media as well as for interdisciplinary researchers interested in the intersection of computer science and journalism.