Research on (verbo-)pictorial metaphors and other rhetorical figures is primarily focused on the genre of advertising, leaving other genres under-investigated. In this study, the authors focus on street art, a visually perceived cross-cultural medium used to address sociopolitical issues. This genre typically combines two interacting semiotic systems – language and depiction – and is thus a form of polysemiotic communication. Their analysis is based on a corpus of 50 street artworks addressing the financial, sociopolitical and migrant/refugee crisis in the city of Athens (2015–2017). They present a data-driven procedure for the identification and interpretation of metaphors and other rhetorical figures in street art, informed by cognitive linguistic and semiotic models. Quantitative analyses show that their models can be reliably applied to street art and can enable them to distinguish metaphors from other rhetorical figures within these images. At the same time, qualitative analyses show that this genre usually requires the integration of conceptual, contextual, socio-cultural and linguistic knowledge in order to achieve successful interpretation of these images. The authors discuss their findings within the theoretical framework of cognitive semiotics.