Complex choices, smart solutions: Bayesian networks in urbanhousing choices

Purpose This study aims to explore the relationship between housing aspirations and sociodemographics of households to be used in sustainable urban planning. With evidence from Istanbul, Ankara and Izmir, it examines how household size, education, income and age influence decisions on the type of dwelling, dwelling size and tenure. Design/methodology/approach The study uses Bayesian Network Analysis (BNA) to represent probabilistic dependencies and to explain direct, as well as indirect, relations between various housing characteristics and household demographic attributes. As opposed to traditional approaches, BNA offers an integrated perspective on housing decisions by unveiling complex interaction patterns that are not captured by standard models. Findings The results indicate that younger, more affluent and more highly educated households favor renting and smaller dwellings, whereas older households favor homeownership and single-family dwellings, often downsizing later in life. Large households favor large housing, whereas car ownership is associated with suburban housing types. Scenario simulations reveal the effect of demographic change on future housing demand. Social implications The findings highlight specific and actionable policy needs. The high demand for rental and more compact housing among young, educated and affluent households, particularly in major cities, necessitates policies that promote high-quality rental housing stock. Furthermore, the choice of older households for smaller apartments in central areas over larger homes underscores the need for “aging in place” solutions. Finally, the strong correlation found between low- and middle-income families, car ownership and larger suburban housing directly demonstrates that the integration of housing and transportation planning is essential to foster inclusive and accessible communities. Originality/value This study adds to the field of housing research through the application of BNA, offering novel insight into the relationship between housing choice and sociodemographics. The scenario analysis and methodological framework give practical suggestions for urban planners and policymakers.