Search for: All records

There has been significant progress in improving the performance of graph neural networks (GNNs) through enhancements in graph data, model architecture design, and training strategies. For fairness in graphs, recent studies achieve fair representations and predictions through either graph data pre-processing (e.g., node feature masking, and topology rewiring) or fair training strategies (e.g., regularization, adversarial debiasing, and fair contrastive learning). How to achieve fairness in graphs from the model architecture perspective is less explored. More importantly, GNNs exhibit worse fairness performance compared to multilayer perception since their model architecture (i.e., neighbor aggregation) amplifies biases. To this end, we aim to achieve fairness via a new GNN architecture. We propose Fair Message Passing (FMP) designed within a unified optimization framework for GNNs. Notably, FMP explicitly renders sensitive attribute usage in forward propagation for node classification task using cross-entropy loss without data pre-processing. In FMP, the aggregation is first adopted to utilize neighbors' information and then the bias mitigation step explicitly pushes demographic group node presentation centers together.In this way, FMP scheme can aggregate useful information from neighbors and mitigate bias to achieve better fairness and prediction tradeoff performance. Experiments on node classification tasks demonstrate that the proposed FMP outperforms several baselines in terms of fairness and accuracy on three real-world datasets. The code is available at https://github.com/zhimengj0326/FMP.

 

Understanding the relationship between spatial structures of cities and environmental hazard exposures is essential for urban health and sustainability planning. However, a critical knowledge gap exists in terms of the extent to which socio-spatial networks shaped by human mobility exacerbate or alleviate urban heat exposures of populations in cities. In this study, we utilize location-based data to construct human mobility networks in twenty metropolitan areas in the U.S. The human mobility networks are analyzed in conjunction with the urban heat characteristics of spatial areas. We identify areas with high and low urban heat exposure and evaluate visitation patterns of populations residing in high and low urban heat areas to other spatial areas with similar and dissimilar urban heat exposure. The results reveal the presence of urban heat traps in the majority of the studied metropolitan areas, wherein populations residing in high-heat exposure areas primarily visited other high-heat exposure zones. Specifically, cities such as Los Angeles, Boston, and Chicago were particularly pronounced as urban heat traps. The results also show a small percentage of human mobility to produce urban heat escalation and heat escapes. The findings from this study provide a better understanding of urban heat exposure in cities based on patterns of human mobility. These findings contribute to a broader understanding of the intersection of human network dynamics and environmental hazard exposures in cities to inform more integrated urban design and planning to promote health and sustainability.

 

In studying resilience in temporal human networks, relying solely on global network measures would be inadequate; latent sub-structural network mechanisms need to be examined to determine the extent of impact and recovery of these networks during perturbations, such as urban flooding. In this study, we utilize high-resolution aggregated location-based data to construct temporal human mobility networks in Houston in the context of the 2017 Hurricane Harvey. We examine motif distribution, motif persistence, temporal stability, and motif attributes to reveal latent sub-structural mechanisms related to the resilience of human mobility networks during disaster-induced perturbations. The results show that urban flood impacts persist in human mobility networks at the sub-structure level for several weeks. The impact extent and recovery duration are heterogeneous across different network types. Also, while perturbation impacts persist at the sub-structure level, global topological network properties indicate that the network has recovered. The findings highlight the importance of examining the microstructures and their dynamic processes and attributes in understanding the resilience of temporal human mobility networks (and other temporal networks). The findings can also provide disaster managers, public officials, and transportation planners with insights to better evaluate impacts and monitor recovery in affected communities.

 

We present a latent characteristic in socio-spatial networks, hazard-exposure heterophily, to capture the extent to which populations with dissimilar hazard exposure could assist each other through social ties. Heterophily is the tendency of unlike individuals to form social ties. Conversely, populations in hazard-prone spatial areas with significant hazard-exposure similarity, homophily, would lack sufficient resourcefulness to aid each other to lessen the impact of hazards. In the context of the Houston metropolitan area, we use Meta’s Social Connectedness data to construct a socio-spatial network in juxtaposition with flood exposure data from National Flood Hazard Layer to analyze flood hazard exposure of spatial areas. The results reveal the extent and spatial variation of hazard-exposure heterophily in the study area. Notably, the results show that lower-income areas have lower hazard-exposure heterophily possibly caused by income segregation and the tendency of affordable housing development to be located in flood zones. Less resourceful social ties in hazard-prone areas due to their high-hazard-exposure homophily may inhibit low-income areas from better coping with hazard impacts and could contribute to their slower recovery. Overall, the results underscore the significance of characterizing hazard-exposure heterophily in socio-spatial networks to reveal community vulnerability and resilience to hazards.

 

Flood nowcasting refers to near-future prediction of flood status as an extreme weather event unfolds to enhance situational awareness. The objective of this study was to adopt and test a novel structured deep-learning model for urban flood nowcasting by integrating physics-based and human-sensed features. We present a new computational modeling framework including an attention-based spatial–temporal graph convolution network (ASTGCN) model and different streams of data that are collected in real-time, preprocessed, and fed into the model to consider spatial and temporal information and dependencies that improve flood nowcasting. The novelty of the computational modeling framework is threefold: first, the model is capable of considering spatial and temporal dependencies in inundation propagation thanks to the spatial and temporal graph convolutional modules; second, it enables capturing the influence of heterogeneous temporal data streams that can signal flooding status, including physics-based features (e.g., rainfall intensity and water elevation) and human-sensed data (e.g., residents’ flood reports and fluctuations of human activity) on flood nowcasting. Third, its attention mechanism enables the model to direct its focus to the most influential features that vary dynamically and influence the flood nowcasting. We show the application of the modeling framework in the context of Harris County, Texas, as the study area and 2017 Hurricane Harvey as the flood event. Three categories of features are used for nowcasting the extent of flood inundation in different census tracts: (i) static features that capture spatial characteristics of various locations and influence their flood status similarity, (ii) physics-based dynamic features that capture changes in hydrodynamic variables, and (iii) heterogeneous human-sensed dynamic features that capture various aspects of residents’ activities that can provide information regarding flood status. Results indicate that the ASTGCN model provides superior performance for nowcasting of urban flood inundation at the census-tract level, with precision 0.808 and recall 0.891, which shows the model performs better compared with other state-of-the-art models. Moreover, ASTGCN model performance improves when heterogeneous dynamic features are added into the model that solely relies on physics-based features, which demonstrates the promise of using heterogenous human-sensed data for flood nowcasting. Given the results of the comparisons of the models, the proposed modeling framework has the potential to be more investigated when more data of historical events are available in order to develop a predictive tool to provide community responders with an enhanced prediction of the flood inundation during urban flood.

 

Lifestyle recovery captures the collective effects of population activities as well as the restoration of infrastructure and business services. This study uses a novel approach to leverage privacy-enhanced location intelligence data, which is anonymized and aggregated, to characterize distinctive lifestyle patterns and to unveil recovery trajectories after 2017 Hurricane Harvey in Harris County, Texas (USA). The analysis integrates multiple data sources to record the number of visits from home census block groups (CBGs) to different points of interest (POIs) in the county during the baseline and disaster periods. For the methodology, the research utilizes unsupervised machine learning and ANOVA statistical testing to characterize the recovery of lifestyles using privacy-enhanced location intelligence data. First, primary clustering using k-means characterized four distinct essential and non-essential lifestyle patterns. For each primary lifestyle cluster, the secondary clustering characterized the impact of the hurricane into four possible recovery trajectories based on the severity of maximum disruption and duration of recovery. The findings further reveal multiple recovery trajectories and durations within each lifestyle cluster, which imply differential recovery rates among similar lifestyles and different demographic groups. The impact of flooding on lifestyle recovery extends beyond the flooded regions, as 59% of CBGs with extreme recovery durations did not have at least 1% of direct flooding impacts. The findings offer a twofold theoretical significance: (1) lifestyle recovery is a critical milestone that needs to be examined, quantified, and monitored in the aftermath of disasters; (2) spatial structures of cities formed by human mobility and distribution of facilities extend the spatial reach of flood impacts on population lifestyles. These provide novel data-driven insights for public officials and emergency managers to examine, measure, and monitor a critical milestone in community recovery trajectory based on the return of lifestyles to normalcy.

 

Community recovery from hazards occurs through various diffusion processes within social and spatial networks of communities. Existing knowledge regarding the diffusion of recovery in community socio-spatial networks, however, is rather limited. To bridge this gap, we created a network diffusion model to characterize the unfolding of population activity recovery in spatial networks of communities. In particular, this study aims to answer the research question “To what extent can the diffusion model capture the spatial patterns of recovery?” Using population activity recovery data derived from location-based information associated with 2017 Hurricane Harvey in the Houston area, we parameterized the threshold-based network diffusion model using the genetic algorithm and then simulated the recovery diffusion process. The results show that the spatial effects of recovery are rather heterogeneous across different areas; some spatial areas demonstrate a greater spatial effect in their recovery. Also, the results show that low-income and minority areas are community recovery multipliers; with faster recovery in these areas corresponding to accelerated recovery for the entire community. Hence, prioritizing these areas in resource allocation during recovery has the potential to accelerate could expedite the recovery of the entire community’s recovery process while promoting recovery equality and equity.

 

This study uses mobility data in the context of 2017 Hurricane Harvey in Harris County to examine the impact of flooding on access to dialysis centers. We examined access dimensions using static and dynamic metrics. The static metric is the shortest distance from census block groups to the closest centers. Dynamic metrics are: 1) redundancy (daily unique number of centers visited), 2) frequency (daily number of visits to dialysis centers), and 3) proximity (visits weighted by distance to dialysis centers). The results show that: the extent of dependence of regions on dialysis centers varies; flooding significantly reduces access redundancy and frequency of dialysis centers; regions with a greater minority percentage and lower household income were likely to experience extensive disruptions; high-income regions more quickly revert to pre-disaster levels; larger centers located in non-flooded areas are critical to absorbing the unmet demand from disrupted facilities. 

Urban flooding disrupts traffic networks, affecting mobility and disrupting residents’ access. Flooding events are predicted to increase due to climate change; therefore, understanding traffic network’s flood-caused disruption is critical to improving emergency planning and city resilience. This study reveals the anatomy of perturbed traffic networks by leveraging high-resolution traffic network data from a major flood event and advanced high-order network analysis. We evaluate travel times between every pairwise junction in the city and assess higher-order network geometry changes in the network to determine flood impacts. The findings show network-wide persistent increased travel times could last for weeks after the flood water has receded, even after modest flood failure. A modest flooding of 1.3% road segments caused 8% temporal expansion of the entire traffic network. The results also show that distant trips would experience a greater percentage increase in travel time. Also, the extent of the increase in travel time does not decay with distance from inundated areas, suggesting that the spatial reach of flood impacts extends beyond flooded areas. The findings of this study provide an important novel understanding of floods’ impacts on the functioning of traffic networks in terms of travel time and traffic network geometry.