This paper explores Airbnb, a peer-to-peer platform for short-term rental of housing accommodation, examining the geographical pattern of those establishments using data from London. Our purpose is to analyse whether or not the diversity of dwelling types correlate with the distribution of listings. We use a measure of spread based on entropy to indicate the diversity of dwelling types and look at its relationship with the distribution of Airbnb establishments, as well as the type of home ownership using correlation analysis. It is important to note that our study only considers domestic building types, and excludes any information on the diversity of land uses. Two important findings emerge from our analysis. Firstly, the spatial location of Airbnb rentals is negatively correlated with the diversity of dwelling types, and positively correlated with a single dwelling type, which corresponds in general to purpose built flats, conversions and flats in commercial buildings. Secondly, Airbnb is associated with areas that have a high proportion of privately rented properties, detracting more than 1.4% of the housing supply into short-term rentals. Such a phenomenon can reach up to 20% in some neighbourhoods, further exacerbating the process of gentrification. Finally, we discuss the implications of these findings as instruments to inform policies associated with the 'sharing' economy in relation to the disruption of the housing structure.