Blended sensory walksPretnar Žagar, Ajda (Avtor) Muršič, Rajko (Avtor) sensory walkingdata scienceimage analyticsquantitative methodsteaching anthropologyanthropology curricula often neglect quantitative and computational methods, leavingstudents underprepared for data-rich research. yet ethnographic practices such as sensorywalks naturally generate digital materials suited for computational analysis. We developed ablended intervention that combines sensory walks with image clustering, introducing datascience while preserving anthropology’s experiential and interpretive focus. Participants usedsmartphones to document sensory impressions during 30–45-minute walks, creating smallpersonal datasets. Images were embedded and clustered with hierarchical clustering. Testedin four workshops with groups ranging from primary school pupils to postgraduate students,the results were explored interactively, interpreted qualitatively, and discussed collectively,linking algorithmic outputs with ethnographic reflection. The activity produced three mainoutcomes. First, clustering highlighted thematic patterns in images, such as contrasts betweenurban and natural motifs. second, participants gained accessible exposure to embeddings,clustering, and exploratory analysis, building data literacy. Third, the exercise prompted criticalreflection on algorithmic methods, raising questions of trust, interpretability, and limits.Participants also reported greater sensory awareness of their surroundings. This approach shows how computational methods can be integrated into anthropology without sacrificinginterpretive depth. Its simplicity, adaptability, and low cost make it a scalable model for teaching and research.20262026-03-17 10:10:56Članek v reviji180813UDK: 004:37:39ISSN pri članku: 0888-4552DOI: 10.1080/08884552.2026.2635502COBISS_ID: 271510787sl