A modification to geographically weighted regression

Assessing the spatial structure of the association between attendance at preschool and children's developmental vulnerabilities in Queensland, Australia

Demographic and educational factors are essential, influential factors of early childhood development. This study aimed to investigate spatial patterns in the association between attendance at preschool and children's developmental vulnerabilities in one or more domain(s) in their first year of full-time school at a small area level in Queensland, Australia. This was achieved by applying geographically weighted regression (GWR) followed by K-means clustering of the regression coefficients. Three distinct geographical clusters were found in Queensland using the GWR coefficients. The first cluster covered more than half of the state of Queensland, including the Greater Brisbane region, and displays a strong negative association between developmental vulnerabilities and attendance at preschool. That is, areas with high proportions of preschool attendance tended to have lower proportions of children with at least one developmental vulnerability in the first year of full-time school. Clusters two and three were characterized by stronger negative associations between developmental vulnerabilities, English as the mother language, and geographic remoteness, respectively. This research provides evidence of the need for collaboration between health and education sectors in specific regions of Queensland to update current service provision policies and to ensure holistic and appropriate care is available to support children with developmental vulnerabilities.

Temporal trend evaluation in monitoring programs with high spatial resolution and low temporal resolution using geographically weighted regression models

Data from monitoring programs with high spatial resolution but low temporal resolution are often overlooked when assessing temporal trends, as the data structure does not permit the use of established trend analysis methods. However, the data include uniquely detailed information about geographically differentiated temporal trends driven by large-scale influences, such as climate or airborne deposition. In this study, we used geographically weighted regression models, extended with a temporal component, to evaluate linear and nonlinear trends in environmental monitoring data. To improve the results, we tested approaches for station-wise pre-processing of data and for validation of the resulting models. To illustrate the method, we used data on changes in total organic carbon (TOC) obtained in a monitoring program of around 4800 Swedish lakes observed once every 6 years between 2008 and 2021. On applying the methods developed here, we identified nonlinear changes in TOC from consistent negative trends over most of Sweden around 2010 to positive trends during later years in parts of the country.

Temporal trend evaluation in monitoring programs with high spatial resolution and low temporal resolution using geographically weighted regression models

Data from monitoring programs with high spatial resolution but low temporal resolution are often overlooked when assessing temporal trends, as the data structure does not permit the use of established trend analysis methods. However, the data include uniquely detailed information about geographically differentiated temporal trends driven by large-scale influences, such as climate or airborne deposition. In this study, we used geographically weighted regression models, extended with a temporal component, to evaluate linear and nonlinear trends in environmental monitoring data. To improve the results, we tested approaches for station-wise pre-processing of data and for validation of the resulting models. To illustrate the method, we used data on changes in total organic carbon (TOC) obtained in a monitoring program of around 4800 Swedish lakes observed once every 6 years between 2008 and 2021. On applying the methods developed here, we identified nonlinear changes in TOC from consistent negative trends over most of Sweden around 2010 to positive trends during later years in parts of the country.

Non-Iterative Multiscale Estimation for Spatial Autoregressive Geographically Weighted Regression Models

Multiscale estimation for geographically weighted regression (GWR) and the related models has attracted much attention due to their superiority. This kind of estimation method will not only improve the accuracy of the coefficient estimators but also reveal the underlying spatial scale of each explanatory variable. However, most of the existing multiscale estimation approaches are backfitting-based iterative procedures that are very time-consuming. To alleviate the computation complexity, we propose in this paper a non-iterative multiscale estimation method and its simplified scenario for spatial autoregressive geographically weighted regression (SARGWR) models, a kind of important GWR-related model that simultaneously takes into account spatial autocorrelation in the response variable and spatial heterogeneity in the regression relationship. In the proposed multiscale estimation methods, the two-stage least-squares (2SLS) based GWR and the local-linear GWR estimators of the regression coefficients with a shrunk bandwidth size are respectively taken to be the initial estimators to obtain the final multiscale estimators of the coefficients without iteration. A simulation study is conducted to assess the performance of the proposed multiscale estimation methods, and the results show that the proposed methods are much more efficient than the backfitting-based estimation procedure. In addition, the proposed methods can also yield accurate coefficient estimators and such variable-specific optimal bandwidth sizes that correctly reflect the underlying spatial scales of the explanatory variables. A real-life example is further provided to demonstrate the applicability of the proposed multiscale estimation methods.

[Determinants of Problem Drinking by Regional Variation among Adult Males in Single-Person Households: Geographically Weighted Regression Model Analysis]

PURPOSE

This study aimed to identify regional differences in problem drinking among adult males in single-person households and predict the determinants.

METHODS

This study used data from the 2019 Community Health Survey. Geographically weighted regression analysis was performed on 8,625 adult males in single-person households who had been consuming alcohol for the past year. The Si-Gun-Gu was selected as the spatial unit.

RESULTS

The top 10 regions for problem drinking among adult males in single-person households were located in the Jeju-do and Jeollanam-do areas near the southern coast, whereas the bottom 10 regions were located in the Incheon and northern Gyeonggi-do areas. Smoking, economic activity, and educational level were common factors affecting problem drinking among this population. Among the determinants of regional disparities in problem drinking among adult males in single-person households, personal factors included age, smoking, depression level, economic activity, educational level, and leisure activity, while regional factors included population and karaoke venue ratio.

CONCLUSION

Problem drinking among adult males in single-person households varies by region, and the variables affecting each particular area differ. Therefore, it is necessary to develop interventions tailored to individuals and regions that reflect the characteristics of each region by prioritizing smoking, economic activity, and educational level as the common factors.

Generalized Geographically Weighted Regression Model within a Modularized Bayesian Framework

Geographically weighted regression (GWR) models handle geographical dependence through a spatially varying coefficient model and have been widely used in applied science, but its general Bayesian extension is unclear because it involves a weighted log-likelihood which does not imply a probability distribution on data. We present a Bayesian GWR model and show that its essence is dealing with partial misspecification of the model. Current modularized Bayesian inference models accommodate partial misspecification from a single component of the model. We extend these models to handle partial misspecification in more than one component of the model, as required for our Bayesian GWR model. Information from the various spatial locations is manipulated via a geographically weighted kernel and the optimal manipulation is chosen according to a Kullback-Leibler (KL) divergence. We justify the model via an information risk minimization approach and show the consistency of the proposed estimator in terms of a geographically weighted KL divergence.

Spatiotemporal clusters and the socioeconomic determinants of COVID-19 in Toronto neighbourhoods, Canada

The aim of this study is to identify spatiotemporal clusters and the socioeconomic drivers of COVID-19 in Toronto. Geographical, epidemiological, and socioeconomic data from the 140 neighbourhoods in Toronto were used in this study. We used local and global Moran's I, and space-time scan statistic to identify spatial and spatiotemporal clusters of COVID-19. We also used global (spatial regression models), and local geographically weighted regression (GWR) and Multiscale Geographically weighted regression (MGWR) models to identify the globally and locally varying socioeconomic drivers of COVID-19. The global regression model identified a lower percentage of educated people and a higher percentage of immigrants in the neighbourhoods as significant predictors of COVID-19. MGWR shows the best fit model to explain the variables affecting COVID-19. The findings imply that a single intervention package for the entire area would not be an effective strategy for controlling COVID-19; a locally adaptable intervention package would be beneficial.

Modeling spatial determinants of initiation of breastfeeding in Ethiopia: A geographically weighted regression analysis

Background

The World Health Organization (WHO) encourages breastfeeding to begin within the first hour after birth in order to save children’s lives. In Ethiopia, different studies are done on the prevalence and determinants of breastfeeding initiation, up to our knowledge, the spatial distribution and the spatial determinants of breast feeding initiation over time are not investigated. Therefore, the objectives of this study were to assess spatial variation and its spatial determinant of delayed initiation of breastfeeding in Ethiopia using Geographically Weighted Regression (GWR).

Methods

A cross-sectional study was undertaken using the nationally representative 2016 Ethiopian Demographic and Health Survey (EDHS) dataset. Global Moran’s I statistic was used to measure whether delayed breastfeeding initiation was dispersed, clustered, or randomly distributed in study area. Ordinary Least Squares (OLS) regression was used to identify factors explaining the geographic variation in delayed breastfeeding initiation. Besides, spatial variability of relationships between dependent and selected predictors was investigated using geographically weighted regression.

Result

A total weighted sample of 4169 children of aged 0 to 23 months was included in this study. Delayed initiation of breastfeeding was spatially varies across the country with a global Moran’s I value of 0.158 at (p-value<0.01). The hotspot (high risk) areas were identified in the Amhara, Afar, and Tigray regions. Orthodox religion, poor wealth index, caesarian section, baby postnatal checkup, and small size of a child at birth were spatially significant factors for delayed breastfeeding initiation in Ethiopia.

Conclusion

In Ethiopia initiation of breastfeeding varies geographically across region. A significant hotspot was identified in the Amhara, Afar, and Tigray regions. The GWR analysis revealed that orthodox religion, poor wealth index, caesarian section, baby postnatal checkup, and small birth weight were spatially significant factors.

Space Syntax in Analysing Bicycle Commuting Routes in Inner Metropolitan Adelaide

Cycling is a particularly favoured for short urban trips because it is a healthy and environmentally benign activity. As a result, urban mobility, quality of life, and public health are enhanced, while traffic congestion and pollution are decreased. In looking beyond the street network in terms of how it affects cyclists’ behavior choices, Bill Hillier’s (1984) outstanding legacy research on spatial space syntax is investigated in this study. The goal of this study is to determine if an urban area’s street network morphology influences commuters’ inclination to ride their bicycles to work. To further understand the nonlinear consequences of street network geometry on the estimation of cycling to work, a logarithmic-transformed regression model that includes base socioeconomic components, urban form, and street network variables represented by space syntax measure factors is developed. In conclusion, this model determined that bike commuting choice is significantly associated with the centrality index of Connectivity, although this is in combination with socioeconomic factors (age, gender, affluence, housing type, and housing price) and built environment factors (share of commercial, educational activities and distance to the CBD) factors. The findings of this study would be of value to planners and policy makers in support of evidence-based policy formulation to improve the design of bicycle networks in suburban regions.

Spatial analysis and geoclimatic factors associated with the incidence of acute lymphoblastic leukemia in Iran during 2006-2014: An environmental epidemiological study

BACKGROUND AND OBJECTIVES

The present study aims to determine the cumulative incidence rate of acute lymphoblastic leukemia (ALL), the degree of spatial autocorrelation and clustering of ALL, the hotspot and coldspots of ALL and geoclimatic conditions affecting the incidence of ALL in Iran and to draw a comparison between global and local regression models.

MATERIALS AND METHODS

In this ecological study, an exploratory-etiologic multiple-group method has been adopted to investigate all children under 15 years of age with ALL in Iran during 2006-2014. Data analysis was performed using Mann Whitney U, Pearson correlation coefficients (PCCs), Global Moran's I, Optimized hotspot analysis (OHSA), Global Poisson regression (GPR), Geographically Weighted Poisson Regression (GWPR) at a significant level of α = 0.05.

RESULTS

The cumulative incidence rate of ALL was estimated at 21,315 per 100,000 Iranian children under 15 years of age. The value of Global Moran's I index was estimated 0.338 and significant (<0.001 P-value). Coldspots were observed in north and northwest of Iran and hotspots were identified in south, southwest and mid-east of Iran. In the present study, Max Temperature of Warmest Month (MTWM) and Direct Normal Irradiation (DNI) were risk factors and Precipitation of the Coldest Quarter (PCQ) and Altitude (AL) were protective factors in the incidence of ALL, even though the non-stationarity of local coefficients and local t-values was clear. GWPR, by capturing and applying spatial heterogeneity and spatial autocorrelation, had a greater performance and goodness of fit than GPR.

DISCUSSION

ALL has created spatial clusters in Iran. The incidence of ALL is the result of synergistic interaction between environmental, infectious, geographical and genetic risk factors. It is recommended to use of local models in ecological studies.

Study of the Urban Heat Island (UHI) Using Remote Sensing Data/Techniques: A Systematic Review

Urban Heat Islands (UHI) consist of the occurrence of higher temperatures in urbanized areas when compared to rural areas. During the warmer seasons, this effect can lead to thermal discomfort, higher energy consumption, and aggravated pollution effects. The application of Remote Sensing (RS) data/techniques using thermal sensors onboard satellites, drones, or aircraft, allow for the estimation of Land Surface Temperature (LST). This article presents a systematic review of publications in Scopus and Web of Science (WOS) on UHI analysis using RS data/techniques and LST, from 2000 to 2020. The selection of articles considered keywords, title, abstract, and when deemed necessary, the full text. The process was conducted by two independent researchers and 579 articles, published in English, were selected. Qualitative and quantitative analyses were performed. Cfa climate areas are the most represented, as the Northern Hemisphere concentrates the most studied areas, especially in Asia (69.94%); Landsat products were the most applied to estimates LST (68.39%) and LULC (55.96%); ArcGIS (30.74%) was most used software for data treatment, and correlation (38.69%) was the most applied statistic technique. There is an increasing number of publications, especially from 2016, and the transversality of UHI studies corroborates the relevance of this topic.

Understanding Spatiotemporal Variations of Ridership by Multiple Taxi Services

Recent years have seen the big growth of app-based taxi services by not only competing for rides with street-hailing taxi services but also generating new taxi rides. Moreover, the innovation in dynamic pricing also makes it competitive in both passenger and driver sides. However, current literature still lacks better understandings of induced changes in spatiotemporal variations in multiple taxi ridership after app-based taxi service launch. This study develops two study cases in New York City to explore impacts of presence of app-based taxi services on daily total and street-hailing taxi rides and impacts of dynamic pricing on hourly app-based taxi rides. Considering the panel data and treatment effect measurement in this problem, we introduce a mixed modeling structure with both geographically weighted panel regression and difference-in-difference estimator. This mixed modeling structure outperforms traditional fixed effects model in our study cases. Empirical analyses identified the significant spatiotemporal variations in impacts of presence of app-based taxi services; for instance, impacts daily total taxi rides in 2014 and 2016 and impacts on street-hailing taxi rides from 2012 to 2016. Moreover, we capture the spatial variations in impacts of dynamic pricing on hourly app-based taxi rides, as well as significant impacts of time of day, day of week, and vehicle supply.

Spatial heterogeneity in the association between particulate matter air pollution and low birth weight in South Korea

As many studies showed the spatial heterogeneity in the association between particulate matter (PM) air pollution and low birth weight (LBW), few studies focused on the variation of local associations at the national scale and related areal characteristics. This study aimed to explore different approaches to estimating local effects of PM with an aerodynamic diameter ≤10 μm (PM₁₀) on LBW across 235 districts in South Korea, to investigate the spatial pattern of local associations, and to examine the relationship with local socio-demographic and environmental characteristics. LBW was identified in 5,692,650 mothers from birth certificate data for 2001-2013. We estimated individual annual-average concentrations of PM₁₀ at centroids of mothers' residential districts by using a previously-validated prediction model. Then, we estimated district-specific odds ratios of LBW for PM₁₀ using modified geographically weighted logistic regression. Here, we applied four approaches with different neighborhood definitions: the distance-based approach within 20- and 40-km bandwidth and the hybrid approach replacing with adjacent districts for urban districts <100 km². In addition, we compared district-specific socioeconomic indicators and emission estimates across three groups of districts that showed significantly positive, no, and significantly negative associations. Medians of district-specific estimates of four approaches were similar to the global estimate and between each other. However, their variability differed with some unreasonably high estimates when a small distance was applied as the neighborhood definition, although spatial pattern was generally similar among the four. The hybrid approach based on the different neighborhood definition by urban and rural areas provided stable risk estimates. Higher risk districts in rural areas were found in more socioeconomically-deprived areas, whereas urban areas showed higher risk districts when their air pollution emissions were higher. Our approach and findings will help identify high risk areas and enhance understanding of geographic determinants.

Clustered spatio-temporal varying coefficient regression model

In regression analysis for spatio-temporal data, identifying clusters of spatial units over time in a regression coefficient could provide insight into the unique relationship between a response and covariates in certain subdomains of space and time windows relative to the background in other parts of the spatial domain and the time period of interest. In this article, we propose a varying coefficient regression method for spatial data repeatedly sampled over time, with heterogeneity in regression coefficients across both space and over time. In particular, we extend a varying coefficient regression model for spatial-only data to spatio-temporal data with flexible temporal patterns. We consider the detection of a potential cylindrical cluster of regression coefficients based on testing whether the regression coefficient is the same or not over the entire spatial domain for each time point. For multiple clusters, we develop a sequential identification approach. We assess the power and identification of known clusters via a simulation study. Our proposed methodology is illustrated by the analysis of a cancer mortality dataset in the Southeast of the U.S.

A retrospective cross-national examination of COVID-19 outbreak in 175 countries: a multiscale geographically weighted regression analysis (January 11-June 28, 2020)

Objective

This study retrospectively examined the health and social determinants of the COVID-19 outbreak in 175 countries from a spatial epidemiological approach.

Methods

We used spatial analysis to examine the cross-national determinants of confirmed cases of COVID-19 based on the World Health Organization official COVID-19 data and the World Bank Indicators of Interest to the COVID-19 outbreak. All models controlled for COVID-19 government measures.

Results

The percentage of the population age between 15-64 years (Age15-64), percentage smokers (SmokTot.), and out-of-pocket expenditure (OOPExp) significantly explained global variation in the current COVID-19 outbreak in 175 countries. The percentage population age group 15-64 and out of pocket expenditure were positively associated with COVID-19. Conversely, the percentage of the total population who smoke was inversely associated with COVID-19 at the global level.

Conclusions

This study is timely and could serve as a potential geospatial guide to developing public health and epidemiological surveillance programs for the outbreak in multiple countries. Removal of catastrophic medical expenditure, smoking cessation, and observing public health guidelines will not only reduce illness related to COVID-19 but also prevent unecessary deaths.

Geographic differences in maternal and child health care utilization in four Ethiopian regions; a cross-sectional study

BACKGROUND

Maternal and child health (MCH) care utilization often vary with geographic location. We analyzed the geographic distribution and determinants of utilization of four or more antenatal care visits, health facility delivery, child immunization, and care utilization for common childhood illnesses across four Ethiopian regions.

METHODS

A cross-sectional community-based study was employed with two-staged stratified cluster sampling in 46 districts of Ethiopia. A total of 6321 women (13-49 years) and 3110 children below the age of 5 years residing in 5714 households were included. We performed a cluster analysis of the selected MCH care utilization using spatial autocorrelation. We identified district-specific relationships between care coverage and selected factors using geocoded district-level data and ordinary least squares and hotspot analysis using Getis Ord Gi*.

RESULTS

Of the 6321women included in the study, 714 had a live birth in the 12 months before the survey. One-third of the women (30, 95% CI 26-34) had made four or more antenatal visits and almost half of the women (47, 95% CI 43-51) had delivered their most recent child at a health facility. Nearly half of the children (48, 95% CI 40-57) with common childhood illnesses (suspected pneumonia, diarrhoea, or fever) sought care at the health facilities. The proportion of fully immunized children was 41% (95%, CI 37-45). Institutional delivery was clustered at district level (spatial autocorrelation, Moron's I = 0.217, P < 0.01). Full immunization coverage was also spatially clustered (Moron's I = 0.156, P-value < 0.1). Four or more antenatal visits were associated with women's age and parity, while the clustering of institutional delivery was associated with the number of antenatal care visits. Clustering of full immunization was associated with household members owning a mobile phone.

CONCLUSIONS

This study showed evidence for geographic clustering in coverage of health facility deliveries and immunization at the district level, but not in the utilization of antenatal care and utilization of health services for common childhood illnesses. Identifying and improving district-level factors that influenced these outcomes may inform efforts to achieve geographical equitability and universal health coverage.

Examining the Local Spatial Variability of Robberies in Saint Louis Using a Multi-Scale Methodology

The current study spatially examines the local variability of robbery rates in the City of Saint Louis, Missouri using both census tract and block group data disaggregated and standardized to the 250- and 500-m raster grid spatial scale. The Spatial Lag Model (SLM) indicated measures of race and stability as globally influencing robbery rates. To explore these relationships further, Geographically Weighted Regression (GWR) was used to determine the local spatial variability. We found that the standardized census tract data appeared to be more powerful in the models, while standardized block group data were more precise. Similarly, the 250-m grid offered greater accuracy, while the 500-m grid was more robust. The GWR models explained the local varying spatial relationships between race and stability and robbery rates in St. Louis better than the global models. The local models indicated that social characteristics occurring at higher-order geographies may influence robbery rates in St. Louis.