Statistical tools that are commonly used to analyse social media networks, such as the way you interact with followers on Twitter or Facebook, inspired myself and colleagues to build an R package for describing how species interact with one another in natural environments. Recently published in Ecology (PDF available above), our article demonstrates how statistical network models, called Markov Random Fields, offer an exciting way to detect biotic interactions. How co-occurring species interact in nature is a crucial mechanism for explaining variation in biodiversity. Network models are increasingly popular in ecology and disease ecology studies, as we recently demonstrated in a global study of helminth parasite sharing in wildlife. Identifying patterns of co-occurrence can be particularly useful for assessing community assembly mechanisms. But since species’ interactions can vary across environmental gradients, they are difficult to detect. Multivariate regression approaches, such as the one we used in earlier research to analyse parasite co-infections, are the current method of choice for analysing co-occurrences. These methods are incredibly useful (a very well-documented example is described by Otso Ovaskainen et al. in Methods in Ecology and Evolution), but they have certain drawbacks that limit our ability to assess biotic interactions. Most importantly, the co-occurrence probabilities of species are often assumed to be static or independent of environmental influences in these models. This is a major drawback in community ecology research.