Information Fusion to Estimate Resilience of Dense Urban Neighborhoods

Diverse sociocultural influences in rapidly growing dense urban areas may induce strain on civil services and reduce the resilience of those areas to exogenous and endogenous shocks. We present a novel approach with foundations in computer and social sciences, to estimate the resilience of dense urban areas at finer spatiotemporal scales compared to the state-of-the-art. We fuse multi-modal data sources to estimate resilience indicators from social science theory and leverage a structured ontology for factor combinations to enhance explainability. Estimates of destabilizing areas can improve the decision-making capabilities of civil governments by identifying critical areas needing increased social services.

💡 Research Summary

The paper addresses the growing challenge of assessing the resilience of densely populated urban neighborhoods, where rapid demographic change and cultural diversity can strain public services and increase vulnerability to both external and internal shocks. Traditional city‑wide statistics are too coarse to capture neighborhood‑level variations, prompting the authors to develop a novel, fine‑grained, data‑driven framework that fuses multiple open‑source data streams with a solid social‑science foundation.

At the core of the methodology is the concept of social capital, a well‑established measure of community health that reflects the degree of trust, reciprocity, and cooperative behavior among residents. The authors distinguish between bonding social capital (strong ties within homogeneous groups) and bridging social capital (weak ties that connect diverse groups), and they incorporate social‑anchor theory, which highlights the role of institutions such as schools, community centers, and places of worship in fostering bridging ties.

To operationalize this abstract construct, the authors build an “Urban Resilience Ontology” that links observable entities—social structures (hospitals, schools, religious buildings, etc.) and social events (markets, festivals, fundraisers)—to two categories of social factors: access (who can reach the facility, distance, cultural or religious barriers) and capacity (the physical size, number of beds, seats, etc.). Each factor receives a weight w ranging from –1 (negative contribution) to +1 (positive contribution), reflecting whether the entity is expected to increase or decrease local social capital.

The empirical case study focuses on Jakarta, Indonesia. Spatial data on the location and type of social structures are extracted from OpenStreetMap, while high‑resolution (100 m × 100 m) population estimates are obtained from WorldPop. The authors then apply Kernel Density Estimation (KDE) to each facility, using a two‑dimensional Gaussian kernel. The amplitude of each kernel is defined as

\