Conference Agenda

Existing research rarely examines how school campus spatial distributions relate to children’s physical activity in the Global South. We recorded the time schoolchildren in four randomly selected public elementary schools in Arapiraca, Brazil spent on moderate-to-vigorous physical activity (MVPA), light physical activity (LPA), and sedentary behavior (SB) using ActiGraph® GT3X+ accelerometers, and derived precise spatial and density ratios from architectural drawings and school records verified with photographs. Age-adjusted multiple regressions indicated that higher indoor-to-outdoor ratios were associated with less MVPA (β = –0.07, p = .040) and LPA (β = –0.08, p = .022) and with more SB (β = 0.12, p < .001). Larger patio-to-indoor ratios were positively associated with MVPA (β = 0.11, p < .001) and LPA (β = 0.10, p = .003), while greater courtyard-to-indoor ratios were associated with more LPA (β = 0.08, p = .042). Density ratios, i.e. schoolchildren per unit area, also had significant correlations: greater total outdoor density with less MVPA (β = –72.62, p = .050), less LPA (β = –84.70, p = .024), and more SB (β = 65.49, p < .001); higher courtyard density with more MVPA (β = 43.76, p = .049) and LPA (β = 51.06, p = .023) and less SB (β = –43.47, p < .001); and greater patio density with less MVPA (β = –42.81, p < .001), more LPA (β = 54.87, p = .024), and less SB (β = –42.81, p < .001). These findings underscore the potential importance of spatial distributions in supporting movement opportunities.

This paper examines how in-situ redevelopment in Indian urban informal housing is associated with women’s everyday social networks, with implications for social support and community resilience. While redevelopment initiatives aim to improve housing quality and infrastructure, they often reconfigure neighborhood layouts and patterns of interaction that have long underpinned women’s informal safety nets. The study investigates how women’s social ties, particularly help networks, differ as residents transition from densely knit informal settlements to vertically structured, formalized housing.

Drawing on structured social network surveys conducted in two urban communities in western India, the study assesses differences in network size and composition, including gender, kinship, age, and neighborhood homophily. Social network mapping techniques are used to compare the diversity and spatial reach of networks in two communities, one waiting to be redeveloped and second already redeveloped. Findings indicate that women in redeveloped housing report smaller, more kin-centric networks and reduced opportunities for informal exchange, whereas residents of informal settlements maintain broader, multilayered ties anchored in shared public and semi-public spaces. These findings could reflect consequences of redevelopment redesign with reduction in incidental interactions and compressing of historically open and gendered social spaces. The paper argues that redevelopment processes, when executed without attention to everyday social infrastructures, risk eroding the relational ecosystems essential for women’s well-being. By foregrounding women’s social networks as a critical lens, this research contributes to debates on gender, urban transformation, and the design of inclusive housing policies in South Asia.

Limited access to quality healthcare remains one of the most pressing global challenges. This issue affects not only rural or crisis-prone areas but also urban populations who face barriers such as provider shortages, affordability, and logistical constraints. In response, telehealth has rapidly expanded, particularly after COVID-19, as a scalable strategy for improving care access. However, the growing reliance on remote services has raised concerns about the quality of care, especially in relation to the nature of patient-provider interaction in non-physical settings.

This study addresses a critical gap by focusing on the often-overlooked influence of spatial and environmental design on virtual healthcare experiences. This qualitative study examines how architectural elements at the room and interface scale influence patient-provider interaction quality across four telehealth modalities and traditional in-person appointment. Simulated sessions included both physical and mental health appointments to capture diverse care contexts. The study involved ten wellness participants and ten healthcare providers, each completing five different appointment conditions, allowing for a rich comparison of environmental experiences across varied scenarios.

The study identified twelve environmental design factors that affect the effectiveness of telehealth spaces. The most impactful factors include camera placement that supports eye contact, acoustic clarity and soundproofing, and ergonomic seating arrangements. Additional variables such as video display quality, visual and functional aids, lighting, background composition, and color were also important. These elements were assessed using post-session interviews, thematic coding, and cross-modality comparison techniques. The findings revealed subtle but meaningful differences in contributing factors between mental and physical health visits, reflecting the unique environmental requirements of each care context. For example, physical health appointments emphasized visual clarity and functional layout, while mental health sessions placed greater weight on emotional comfort and privacy.

These findings provide evidence-based design criteria that can be applied to a variety of telehealth environments, from hospital rooms and provider offices to community-based telehealth hubs. Designers and healthcare planners can use these insights to optimize spatial conditions that support presence, empathy, and clarity in virtual care delivery. The results are globally relevant and offer guidance for strengthening care delivery, particularly in underserved areas and during public health crises. These findings contribute to key theoretical frameworks, including social presence theory, cognitive load theory, and biophilic design theory, and position environmental design as essential to achieving equitable, high-quality telehealth care. By integrating spatial awareness into telehealth infrastructure planning, this work reinforces the role of architectural research in addressing systemic healthcare inequities through locally informed, design-driven solutions.

Accurate prediction of overall thermal comfort (OTC) in radiant heating systems is essential for creating energy-efficient and user-centered environments. Radiant heating systems have emerged as a promising heating system that can provide uniform thermal comfort for building occupants by reducing temperature asymmetries in indoor environments. This study investigates the application of machine learning (ML) to predict OTC using experimental data on local body parts' thermal comfort. Measurements of local body parts’ skin temperatures and subjective comfort votes were collected using radiant heating devices. Traditional indices, such as Predicted Mean Vote (PMV), often fail to account for localized discomfort caused by non-uniform radiant heating. This work builds on prior chamber experiments to present an advanced machine-learning (ML) framework for predicting overall thermal sensation (o_sensation) and overall thermal comfort (o_comfort) from local physiological and perceptual inputs collected under directed radiant heating. Leveraging an expanded dataset of 100,000 samples produced from the KIT LOBSTER chamber experiments (original pilot ≈12,000 samples), we evaluated seven ML algorithms—Random Forest, XGBoost, LightGBM, CatBoost, Gradient Boosting, Ridge Regression, and K-Nearest Neighbors—within a multi-output architecture. Models were assessed with MAE, RMSE, and R² and interpreted through node–performance maps, normalized radar charts, network graphs, histograms, boxplots, and feature-importance analysis. Tree-based ensemble models, particularly Random Forest and XGBoost, consistently outperformed linear and instance-based methods, showing tighter error distributions and higher explanatory power. o_sensation proved more tightly coupled to skin temperature signals than o_comfort, which exhibited higher variance likely due to cognitive and contextual influences. We discuss practical implications for adaptive radiant HVAC control, limitations due to chamber conditions and data modalities, and future directions including multimodal sensing, temporal modeling, and validation. The integration of physiological and environmental parameters offers a personalized approach, bridging gaps in traditional methods.