Tweeting migration: a decade of shifting sentiment across local areas of Great Britain

Public attitudes towards migration shape politics, social cohesion and policy, yet existing evidence relies largely on surveys that cannot resolve fine-scale geographic variation. Here, we use 925,000 geolocated migration-related tweets collected across Great Britain between 2013 and 2022 to produce the first decade-long, district-level analysis of migration sentiment derived from digital trace data. We show that migration sentiment exhibits pronounced local spatial heterogeneity that is obscured by conventional survey approaches. Areas with larger migrant populations consistently express more positive sentiment, supporting contact-based explanations of attitude formation. By contrast, recent migrant inflows show no measurable association with negative sentiment, providing little evidence for local group-threat dynamics. Apparent urban-rural differences disappear once demographic composition is taken into account, indicating that spatial variation primarily reflects differences in the concentration of younger and more highly educated populations rather than urban context itself. Longitudinal analyses further reveal increasing polarisation in online migration discourse over the decade despite broad shifts in average sentiment. Together, our findings demonstrate how digital trace data can complement traditional surveys by providing temporally rich and geographically granular evidence on public opinion, offering new opportunities to evaluate theories of migration attitudes and to monitor local social responses to migration in near real time.