Employing urban planning and design to promote active travel modes, such as walking and biking, are important for decarbonizing urban mobility. This paper proposes a modeling framework that investigates the interdependencies between the built environment and travel behaviors through the lens of urban microclimates. We combine travel data, built environment data, and Universal Thermal Climate Index (UTCI) calculations for New York City and train a predictive model for intra-city mobility patterns. Then, impacts of UTCI features are studied through a sensitivity analysis and a spatial heterogeneity analysis. Results show that impacts of UTCI features can account for up to 4% change in the choice of active travel mode in dense urban areas. Also, impacts of microclimate vary across different travel contexts regarding season, time of day, activity, built environment, and traveler type. Our methodology and findings can inform future decision-making of microclimate-oriented spatial planning and design interventions.