This article looks into using large-scale social media data in SSH research and in particular in studies of language variation and change. It presents a study that investigates the role of social networks in linguistic variability. Previous studies have convincingly shown that networks in which people are connected to each other in loose ties tend to contribute positively to language change. Conversely, networks in which people are closely connected to each other inhibit change. This conclusion is, however, based on small datasets from small networks, and this study tests if the difference is diluted when network size is closer to human average. The results from 3,935 networks suggest this to be the case. Towards the end, the article suggests numerous ways in which large-scale social media data and the use of data intensive methodologies could be increased and encouraged in SSH research.