Learning Social Meta-knowledge for Nowcasting Human Mobility in Disaster

Human mobility nowcasting is a fundamental research problem for intelligent transportation planning, disaster responses and management, etc. In particular, human mobility under big disasters such as hurricanes and pandemics deviates from its daily routine to a large extent, which makes the task more challenging. Existing works mainly focus on traffic or crowd flow prediction in normal situations. To tackle this problem, in this study, disaster-related Twitter data is incorporated as a covariate to understand the public awareness and attention about the disaster events and thus perceive their impacts on the human mobility. Accordingly, we propose a Meta-knowledge-Memorizable Spatio-Temporal Network (MemeSTN), which leverages memory network and meta-learning to fuse social media and human mobility data. Extensive experiments over three real-world disasters including Japan 2019 typhoon season, Japan 2020 COVID-19 pandemic, and US 2019 hurricane season were conducted to illustrate the effectiveness of our proposed solution. Compared to the state-of-the-art spatio-temporal deep models and multivariate-time-series deep models, our model can achieve superior performance for nowcasting human mobility in disaster situations at both country level and state level. © 2023 ACM.

Spatio-temporal dynamic of the COVID-19 epidemic and the impact of imported cases in Rwanda.

INTRODUCTION: Africa was threatened by the coronavirus disease 2019 (COVID-19) due to the limited health care infrastructure. Rwanda has consistently used non-pharmaceutical strategies, such as lockdown, curfew, and enforcement of prevention measures to control the spread of COVID-19. Despite the mitigation measures taken, the country has faced a series of outbreaks in 2020 and 2021. In this paper, we investigate the nature of epidemic phenomena in Rwanda and the impact of imported cases on the spread of COVID-19 using endemic-epidemic spatio-temporal models. Our study provides a framework for understanding the dynamics of the epidemic in Rwanda and monitoring its phenomena to inform public health decision-makers for timely and targeted interventions. RESULTS: The findings provide insights into the effects of lockdown and imported infections in Rwanda's COVID-19 outbreaks. The findings showed that imported infections are dominated by locally transmitted cases. The high incidence was predominant in urban areas and at the borders of Rwanda with its neighboring countries. The inter-district spread of COVID-19 was very limited due to mitigation measures taken in Rwanda. CONCLUSION: The study recommends using evidence-based decisions in the management of epidemics and integrating statistical models in the analytics component of the health information system.

Spatio-Temporal Bayesian Models for Malaria Risk Using Survey and Health Facility Routine Data in Rwanda.

INTRODUCTION: Malaria is a life-threatening disease ocuring mainly in developing countries. Almost half of the world's population was at risk of malaria in 2020. Children under five years age are among the population groups at considerably higher risk of contracting malaria and developing severe disease. Most countries use Demographic and Health Survey (DHS) data for health programs and evaluation. However, malaria elimination strategies require a real-time, locally-tailored response based on malaria risk estimates at the lowest administrative levels. In this paper, we propose a two-step modeling framework using survey and routine data to improve estimates of malaria risk incidence in small areas and enable quantifying malaria trends. METHODS: To improve estimates, we suggest an alternative approach to modeling malaria relative risk by combining information from survey and routine data through Bayesian spatio-temporal models. We model malaria risk using two steps: (1) fitting a binomial model to the survey data, and (2) extracting fitted values and using them in the Poison model as nonlinear effects in the routine data. We modeled malaria relative risk among under-five-year old children in Rwanda. RESULTS: The estimation of malaria prevalence among children who are under five years old using Rwanda demographic and health survey data for the years 2019-2020 alone showed a higher prevalence in the southwest, central, and northeast of Rwanda than the rest of the country. Combining with routine health facility data, we detected clusters that were undetected based on the survey data alone. The proposed approach enabled spatial and temporal trend effect estimation of relative risk in local/small areas in Rwanda. CONCLUSIONS: The findings of this analysis suggest that using DHS combined with routine health services data for active malaria surveillance may provide provide more precise estimates of the malaria burden, which can be used toward malaria elimination targets. We compared findings from geostatistical modeling of malaria prevalence among under-five-year old children using DHS 2019-2020 and findings from malaria relative risk spatio-temporal modeling using both DHS survey 2019-2020 and health facility routine data. The strength of routinely collected data at small scales and high-quality data from the survey contributed to a better understanding of the malaria relative risk at the subnational level in Rwanda.

A mechanistic spatio-temporal modeling of COVID-19 data.

Understanding the evolution of an epidemic is essential to implement timely and efficient preventive measures. The availability of epidemiological data at a fine spatio-temporal scale is both novel and highly useful in this regard. Indeed, having geocoded data at the case level opens the door to analyze the spread of the disease on an individual basis, allowing the detection of specific outbreaks or, in general, of some interactions between cases that are not observable if aggregated data are used. Point processes are the natural tool to perform such analyses. We analyze a spatio-temporal point pattern of Coronavirus disease 2019 (COVID-19) cases detected in Valencia (Spain) during the first 11 months (February 2020 to January 2021) of the pandemic. In particular, we propose a mechanistic spatio-temporal model for the first-order intensity function of the point process. This model includes separate estimates of the overall temporal and spatial intensities of the model and a spatio-temporal interaction term. For the latter, while similar studies have considered different forms of this term solely based on the physical distances between the events, we have also incorporated mobility data to better capture the characteristics of human populations. The results suggest that there has only been a mild level of spatio-temporal interaction between cases in the study area, which to a large extent corresponds to people living in the same residential location. Extending our proposed model to larger areas could help us gain knowledge on the propagation of COVID-19 across cities with high mobility levels.

ST-COVID: a Deep Multi-View Spatio-temporal Model for COVID-19 Forecasting

The outbreak of COVID-19 has caused a dramatic loss of human life worldwide. Reliable prediction results are crucial on pandemic prevention and control in the early stage. However, it is a very challenging task due to insufficient data and dynamic virus spread pattern. Unlike most existing works only considering local data for a given region, we propose a spatio-temporal prediction model (ST-COVID) for COVID-19 forecasting to borrow experience from historical observations of other regions. Specifically, our proposed model consists of two views: spatial view (modeling global spatial connectivity with neighbor regions in geography and semantic space via GCNs), temporal view (extracting local and global latent temporal trend via CNNs and GRU). Extensive experiments on two real-world datasets at state and county level in US indicate that the proposed model outperforms over nine baselines in both short-term and long-term prediction. © 2022 IEEE.

Study on Urban Travel Volume During the Outbreak of COVID-19

The distribution and change of travel intensity reflect the pattern of the city and the activity of trip population. It is important to understand the pattern of the city and the activity of trip flow for urban planning and government decision-making. This paper constructs a Bayesian hierarchical spatiotemporal model with three effects: space, time, and space-time, which uses the travel intensity data during the outbreak of the novel coronavirus (COVID-19) in Hubei province (2020.01.01–2020.05.02). With the help of Markoff’s Monte Carlo method, this paper analyzes the distribution and fluctuation of traffic flow in each city of Hubei province. The results show that the space-time model does not deteriorate compared with the main space model. The study found that nearly 41% of cities with a spatial effect higher than 1 were active during the epidemic in Hubei province and the time effect of travel intensity in Hubei province dropped rapidly from 2 to 0.5 after cities in Hubei province issued measures to close the cities one after another, which lasted nearly a month. Strict social distance intervention is one of the important reasons for Hubei province to control the epidemic effectively in a few months. At the same time, in the stability analysis of the city, we found that Wuhan belongs to an unstable area, which is unfavorable to the control of COVID-19. The research results provide a certain perspective for COVID-19 prevention and control: when there are confirmed patients in the province, we believe that the government should first pay attention to those cities with high spatial effect and instability. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Predicting the COVID-19 in the Metropolitan Region (Chile) using a GCN-LSTM neural network

COVID-19 is considered one of the largest pandemics in recent times. Predicting the number of future COVID-19 cases is extremely important for governments in order to make decisions about mobility restrictions, and for hospitals to be able to manage medical supplies, as well as health staff. Most of the predictions of COVID-19 cases are based on mathematical-epidemiological models such as the SEIR and SIR models. In our work, we propose a model of neural networks GCN-LSTM (Graph Convolutional Network - Long Short Term Memory) to predict the spatio-temporal rate incidence of COVID-19 in the Metropolitana Region, Chile. While the GCN network incorporates the spatial correlation in the nearby municipalities, the LSTM network considers the temporal correlation for the prediction over time. To interpolate the missing daily data for the network input, the use of the GAM (Generalized Additive Model) model is proposed. The results show better predictions for some municipalities with higher habitat density. © 2021 IEEE.

Early Detection of COVID-19 Hotspots Using Spatio-Temporal Data

Recently, the Centers for Disease Control and Prevention (CDC) has worked with other federal agencies to identify counties with increasing coronavirus disease 2019 (COVID-19) incidence (hotspots) and offers support to local health departments to limit the spread of the disease. Understanding the spatio-temporal dynamics of hotspot events is of great importance to support policy decisions and prevent large-scale outbreaks. This paper presents a spatio-temporal Bayesian framework for early detection of COVID-19 hotspots (at the county level) in the United States. We assume both the observed number of cases and hotspots depend on a class of latent random variables, which encode the underlying spatio-temporal dynamics of the transmission of COVID-19. Such latent variables follow a zero-mean Gaussian process, whose covariance is specified by a non-stationary kernel function. The most salient feature of our kernel function is that deep neural networks are introduced to enhance the model's representative power while still enjoying the interpretability of the kernel. We derive a sparse model and fit the model using a variational learning strategy to circumvent the computational intractability for large data sets. Our model demonstrates better interpretability and superior hotspot-detection performance compared to other baseline methods. IEEE

The Theory Prospect of Crowd Dynamics-oriented Observation

During the development of COVID-19 virus's global epidemic, the fundamental research and various applications of crowd dynamics-oriented observation theories have attracted much attention from many researchers and people all over the world within some related disciplines, such as public health, clinical medicine, geography, public management, etc. Researchers conducted many interdisciplinary explorations in theories and methods of monitoring epidemic dynamics scientifically, preventing and controlling spatial transmission precisely, predicting accurately, and responding effectively. However, no crowd dynamics-oriented observation theories have been proposed in literature so far. This paper revisits the concept and introduces a theory framework of crowd dynamics-oriented observation, which tries to include the core theories of observation from geospatial big data and to support diverse potential developments. Firstly, this article introduces the research background of crowd dynamics-oriented observation, and then summarizes its three core questions (how to observe its change, how to analyze its change, and how to control its change). From the interdiscipline view of geographic information science, surveying and mapping science, this paper explains the research significance and disciplinary value of crowd dynamics-oriented observation theories. Secondly, this paper introduces a framework of crowd dynamics-oriented observation and its spatiotemporal application, and then elaborates on the bottleneck problems of the key observation theories of crowd dynamics, such as fundamental space-time framework theory, space-time quantification and comprehensive observation theory, spatiotemporal process optimization theory, etc. Thirdly, this paper preliminarily introduces some changes of crowd dynamics-oriented observation theories, for example, refined observation driven by the application needs of digital society governance and public safety/health emergency, personal privacy protection and personalized observations by balancing the public interest and personal privacies, the development of integrated observation theories for human-oriented observation and earth-oriented observation, and the theory of crowd dynamicsoriented observation for high-level management and service. Finally, this article points out the potential directions of crowd dynamics-oriented observation theory and methods, such as, the development of big datadriven crowd perception, multi-space refined crowd dynamics observation, and human-land systematical interaction modeling, so as to realize some differentiated, integrated, and hierarchical crowd dynamics-oriented observations. All potential theories are helpful to the scientific decision-making of public management and public service. The crowd dynamics-oriented observation theory should focus on the fundamental research questions related to studying, analyzing, and servicing human beings, which has become a research frontier in geospatial information science, and could play very important roles in supporting national development strategies, such as "New urbanization", "beautiful China", "artificial intelligence", and "new infrastructure", so as to contribute to a green, efficient, smart, and sustainable regional and urban development. © 2021, Science Press. All right reserved.

A comparison of multiple neighborhood matrix specifications for spatio-temporal model fitting: a case study on COVID-19 data.

Establishing proper neighbor relations between a set of spatial units under analysis is essential when carrying out a spatial or spatio-temporal analysis. However, it is usual that researchers choose some of the most typical (and simple) neighborhood structures, such as the first-order contiguity matrix, without exploring other options. In this paper, we compare the performance of different neighborhood matrices in the context of modeling the weekly relative risk of COVID-19 over small areas located in or near Valencia, Spain. Specifically, we construct contiguity-based, distance-based, covariate-based (considering mobility flows and sociodemographic characteristics), and hybrid neighborhood matrices. We evaluate the goodness of fit, the overall predictive quality, the ability to detect high-risk spatio-temporal units, the capability to capture the spatio-temporal autocorrelation in the data, and the goodness of smoothing for a set of spatio-temporal models based on each of the neighborhood matrices. The results show that contiguity-based matrices, some of the distance-based matrices, and those based on sociodemographic characteristics perform better than the matrices based on k-nearest neighbors and those involving mobility flows. In addition, we test the linear combination of some of the constructed neighborhood matrices and the reweighting of these matrices after eliminating weak neighbor relations, without any model improvement.

Delay in the Effect of Restricting Community Mobility on the Spread of COVID-19 During the First Wave in the United States.

BACKGROUND: It remains unclear how changes in human mobility shaped the transmission dynamic of coronavirus disease 2019 (COVID-19) during its first wave in the United States. METHODS: By coupling a Bayesian hierarchical spatiotemporal model with reported case data and Google mobility data at the county level, we found that changes in movement were associated with notable changes in reported COVID-19 incidence rates about 5 to 7 weeks later. RESULTS: Among all movement types, residential stay was the most influential driver of COVID-19 incidence rate, with a 10% increase 7 weeks ago reducing the disease incidence rate by 13% (95% credible interval, 6%-20%). A 10% increase in movement from home to workplaces, retail and recreation stores, public transit, grocery stores, and pharmacies 7 weeks ago was associated with an increase of 5%-8% in the COVID-10 incidence rate. In contrast, parks-related movement showed minimal impact. CONCLUSIONS: Policy-makers should anticipate such a delay when planning intervention strategies restricting human movement.

Endemic-epidemic models to understand COVID-19 spatio-temporal evolution.

We propose an endemic-epidemic model: a negative binomial space-time autoregression, which can be employed to monitor the contagion dynamics of the COVID-19 pandemic, both in time and in space. The model is exemplified through an empirical analysis of the provinces of northern Italy, heavily affected by the pandemic and characterized by similar non-pharmaceutical policy interventions.

A Bayesian spatio-temporal analysis of mortality rates in Spain: application to the COVID-19 2020 outbreak.

BACKGROUND: The number of deaths attributable to COVID-19 in Spain has been highly controversial since it is problematic to tell apart deaths having COVID as the main cause from those provoked by the aggravation by the viral infection of other underlying health problems. In addition, overburdening of health system led to an increase in mortality due to the scarcity of adequate medical care, at the same time confinement measures could have contributed to the decrease in mortality from certain causes. Our aim is to compare the number of deaths observed in 2020 with the projection for the same period obtained from a sequence of previous years. Thus, this computed mortality excess could be considered as the real impact of the COVID-19 on the mortality rates. METHODS: The population was split into four age groups, namely: (< 50; 50-64; 65-74; 75 and over). For each one, a projection of the death numbers for the year 2020, based on the interval 2008-2020, was estimated using a Bayesian spatio-temporal model. In each one, spatial, sex, and year effects were included. In addition, a specific effect of the year 2020 was added ("outbreak"). Finally, the excess deaths in year 2020 were estimated as the count of observed deaths minus those projected. RESULTS: The projected death number for 2020 was 426,970 people, the actual count being 499,104; thus, the total excess of deaths was 72,134. However, this increase was very unequally distributed over the Spanish regions. CONCLUSION: Bayesian spatio-temporal models have proved to be a useful tool for estimating the impact of COVID-19 on mortality in Spain in 2020, making it possible to assess how the disease has affected different age groups accounting for effects of sex, spatial variation between regions and time trend over the last few years.

The Signature Features of COVID-19 Pandemic in a Hybrid Mathematical Model-Implications for Optimal Work-School Lockdown Policy.

The new COVID-19 pandemic has challenged policymakers on key issues. Most countries have adopted "lockdown" policies to reduce the spatial spread of COVID-19, but they have damaged the economic and moral fabric of society. Mathematical modeling in non-pharmaceutical intervention policy management has proven to be a major weapon in this fight due to the lack of an effective COVID-19 vaccine. A new hybrid model for COVID-19 dynamics using both an age-structured mathematical model based on the SIRD model and spatio-temporal model in silico is presented, analyzing the data of COVID-19 in Israel. Using the hybrid model, a method for estimating the reproduction number of an epidemic in real-time from the data of daily notification of cases is introduced. The results of the proposed model are confirmed by the Israeli Lockdown experience with a mean square error of 0.205 over 2 weeks. The use of mathematical models promises to reduce the uncertainty in the choice of "Lockdown" policies. The unique use of contact details from 2 classes (children and adults), the interaction of populations depending on the time of day, and several physical locations, allow a new look at the differential dynamics of the spread and control of infection.

The impact of modelling choices on modelling outcomes: a spatio-temporal study of the association between COVID-19 spread and environmental conditions in Catalonia (Spain).

The choices that researchers make while conducting a statistical analysis usually have a notable impact on the results. This fact has become evident in the ongoing research of the association between the environment and the evolution of the coronavirus disease 2019 (COVID-19) pandemic, in light of the hundreds of contradictory studies that have already been published on this issue in just a few months. In this paper, a COVID-19 dataset containing the number of daily cases registered in the regions of Catalonia (Spain) since the start of the pandemic to the end of August 2020 is analysed using statistical models of diverse levels of complexity. Specifically, the possible effect of several environmental variables (solar exposure, mean temperature, and wind speed) on the number of cases is assessed. Thus, the first objective of the paper is to show how the choice of a certain type of statistical model to conduct the analysis can have a severe impact on the associations that are inferred between the covariates and the response variable. Secondly, it is shown how the use of spatio-temporal models accounting for the nature of the data allows understanding the evolution of the pandemic in space and time. The results suggest that even though the models fitted to the data correctly capture the evolution of COVID-19 in space and time, determining whether there is an association between the spread of the pandemic and certain environmental conditions is complex, as it is severely affected by the choice of the model.

Effects of long-term exposure to air pollutants on the spatial spread of COVID-19 in Catalonia, Spain.

BACKGROUND: The risk of infection and death by COVID-19 could be associated with a heterogeneous distribution at a small area level of environmental, socioeconomic and demographic factors. Our objective was to investigate, at a small area level, whether long-term exposure to air pollutants increased the risk of COVID-19 incidence and death in Catalonia, Spain, controlling for socioeconomic and demographic factors. METHODS: We used a mixed longitudinal ecological design with the study population consisting of small areas in Catalonia for the period February 25 to May 16, 2020. We estimated Generalized Linear Mixed models in which we controlled for a wide range of observed and unobserved confounders as well as spatial and temporal dependence. RESULTS: We have found that long-term exposure to nitrogen dioxide (NO2) and, to a lesser extent, to coarse particles (PM10) have been independent predictors of the spatial spread of COVID-19. For every 1 µm/m3 above the mean the risk of a positive test case increased by 2.7% (95% credibility interval, ICr: 0.8%, 4.7%) for NO2 and 3.0% (95% ICr: -1.4%,7.44%) for PM10. Regions with levels of NO2 exposure in the third and fourth quartile had 28.8% and 35.7% greater risk of a death, respectively, than regions located in the first two quartiles. CONCLUSION: Although it is possible that there are biological mechanisms that explain, at least partially, the association between long-term exposure to air pollutants and COVID-19, we hypothesize that the spatial spread of COVID-19 in Catalonia is attributed to the different ease with which some people, the hosts of the virus, have infected others. That facility depends on the heterogeneous distribution at a small area level of variables such as population density, poor housing and the mobility of its residents, for which exposure to pollutants has been a surrogate.