Providing equitable and more sustainable access to basic amenities is key to cutting carbon emissions and increasing social equity in cities. This paper applies a data-driven approach to generate an activity-based accessibility index for British cities. To do so, we employ a two-step approach. Using a self-organising map, we first generate behavioural profiles based on the British Time Use Survey. We then use the resulting clusters to determine the weight of the different amenities in our accessibility index. Our preliminary results seem to provide a promising avenue for applying data-driven methods to generate weighted accessibility scores. Future work will incorporate Census 2021 data in order to better characterise neighbourhoods in terms of both accessibility and their demographic characteristics.