Geographically Weighted Poisson Regression Research Articles

Exploring the relationship between built environment and bike-sharing demand: Does the trip length matter?

ABSTRACT Over the past decade, bike-sharing has garnered significant attention in both research and practice. As a manpower-driven transportation mode, the usage of bikes seems more sensitive to trip length, since one could take a shared bike to a destination where is too far to walk, or choose it for simply replacing walking when going to a nearby place. This paper identifies a threshold of bike-sharing trip lengths from bike-sharing trace data, and employs the Semiparametric Geographically Weighted Poisson Regression model to investigate the relationship between built environment and bike-sharing demand with different lengths by considering the heterogeneity in the relationship. Results show that built environment has heterogeneous effects on the bike-sharing demand in urban areas, and the effects differ across groups with trip lengths. The findings contribute to understanding the relationships between built environment and bike-sharing demand, and providing support for the placements and dispatching of shared bikes.

Analyzing the impact of planning schemes on the neighborhood-scale PM2.5 distribution through Pix2PixGAN and crowd-source data

The distribution of PM2.5 concentrations at the neighborhood scale is interfered with not only by spatial morphology, but also by spatial functions, and this effect shows heterogeneity in different urban locations due to differences in spatial dynamics. Taking the area within the Fourth Ring Road of Beijing in 2020 as an example, this paper provides urban planners with an evaluation framework for fine-grained prediction of neighborhood-scale PM2.5 distribution based on the Pix2PixGAN model and the multi-source data with crowd-source data at its core. Through the crowd-source data, the spatial functions and corresponding spatial dynamics in the neighborhood scale can be characterized in detail. Then the framework utilizes Geographically weighted poisson regression (GWPR) to screen spatial functions with significant geographic correlation with PM2.5 distribution and their heterogeneity coefficients. Based on the filtered functions, heterogeneity coefficients, and PM2.5 distribution raster, the dataset of Pix2PixGAN model is composed, and the optimal model with the most accurate PM2.5 distribution prediction is trained and selected. Finally, the paper illustrates how to analyze the qualitative and quantitative impacts of planning proposals on the distribution of PM2.5 through the optimal model, using six planning scenarios as examples.

Geographically Weighted Poisson Regression for Modeling the Number of Maternal Deaths in Papua Province

Introduction/Main Objectives: Maternal Mortality Rate (MMR) in Indonesia is one of the main focuses in achieving the third Sustainable Development Goals (SDGs) in 2030. Background Problems: The Central Statistics Agency states that the MMR in Papua Province is the highest, reaching 565. Novelty: Given the diverse geographical conditions of each district/city in Papua Province, an analysis was carried out. Research Methods: Using the Geographically Weighted Poisson Regression (GWPR) method with the response variable being maternal mortality rates and variables predictors of health, social, and environmental factors. Finding/Results: Fixed Gaussian kernel GWPR is the best model with an AIC value of 27.6. Variable significantly influencing MMR include the percentage of households with access to adequate sanitation, the number of recipients of food assistance programs, and the number of doctors.

RAINFALL MODELING USING THE GEOGRAPHICALLY WEIGHTED POISSON REGRESSION METHOD

Rainfall is an important parameter in understanding the climate and environment in the Malang Regency area. This research aims to model the distribution of rainfall in this region using the Geographically Weighted Poisson Regression (GWPR) method. GWPR is a spatial statistical approach that allows us to understand changes in inhomogeneous rainfall patterns throughout the Malang Regency area. Rainfall data collected from weather stations over several years was used in this study. We use GWR to study the relationship between various environmental factors, such as topography, vegetation, and land use, and rainfall distribution in Malang Regency. The results of the GWR analysis provide a deeper understanding of the spatial differences in the influence of these factors on rainfall. By applying GWR, we can find out how certain factors contribute to different rainfall patterns in certain regions. Rainfall modeling using the Geographically Weighted Poisson Regression (GWPR) method combines the power of Poisson regression in analyzing calculated data with the advantages of GWR in modeling spatial variability. GWPR allows us to identify and map rainfall distribution patterns that vary in geographic space. The main advantage of GWPR is its ability to provide local adjustments and capture the spatial variability associated with rainfall distribution. The results of the modeling analysis show that the GWPR is better, marked by the smallest AIC value, namely 336.84, compared to the generalized poisson regression model, namely 337.76.

MODELING FACTORS AFFECTING EDUCATED UNEMPLOYMENT ON JAVA ISLAND USING GEOGRAPHICALLY WEIGHTED POISSON REGRESSION MODEL

The eighth goal of the SDGs, which aim to promote sustained, inclusive and sustainable economic growth, full and productive employment and decent work for all, addresses the problem of unemployment. Indonesia, the fourth-largest country in the world, keeps on dealing with unemployment and its negative consequences. Three provinces on the island of Java have higher unemployment rates for educated people than any other provinces. The purpose of this study is to examine the variables affecting educated unemployment in Java. This study uses cross-sectional data published from BPS-Statistics Indonesia website and the Indonesia Investment Coordinating Board (BKPM) for 119 regencies/cities across six provinces on Java Island in 2021. The predictor variables are Gross Regional Domestic Product (GRDP), investment, labor force participation rate, mean years of schooling, regency/city minimum wage, and inflation. The number of working-age population is used as an exposure measure. Four models were used to analyze the factors affecting educated unemployment on Java Island: Poisson regression model and Geographically Weighted Poisson Regression (GWPR) model, both with and without an exposure. Based on the smallest AIC/AICc, the best model is the GWPR model with an exposure. This model creates 11 groups of locations based on the same predictor variables that significantly affect educated unemployment

Geographically Weighted Zero-Inflated Negative Binomial Regression: A general case for count data

Poisson and Negative Binomial Regression Models are often used to describe the relationship between a count dependent variable and a set of independent variables. However, these models fail to analyze data with an excess of zeros, being Zero-Inflated Poisson (ZIP) and Zero-Inflated Negative Binomial (ZINB) models the most appropriate to fit this kind of data. To Incorporate the spatial dimension into the count data models, Geographically Weighted Poisson Regression (GWPR), Geographically Weighted Negative Binomial Regression (GWNBR) and Geographically Weighted Zero-Inflated Poisson Regression (GWZIPR) have been developed, but the zero-inflation part of the negative binomial distribution is undeveloped in order to incorporate the overdispersion and the excess of zeros, as was at the beginning of the COVID-19 pandemic, whereas some places were having an outbreak of cases and in others places, there were no cases yet. Therefore, we propose a Geographically Weighted Zero-Inflated Negative Binomial Regression (GWZINBR) model which can be considered a general case for count data, since locally it can become a GWZIPR, GWNBR or a GWPR model. We applied this model to simulated data and to the cases of COVID-19 in South Korea at the beginning of the pandemic in 2020 and the results showed a better understanding of the phenomenon compared to the GWNBR model.

Göç Sonrası Uyum Sürecı̇nde İş Rolünün Coğrafi Ağırlıklı Poisson Regresyon İle İncelenmesi: Van Örneği

Migration is a process of social change that involves the geographical relocation of people from one settlement to another, either permanently or temporarily, in order to spend all or part of their future lives. Van province is among the provinces receiving migration due to its geopolitical location and level of development. Migrant individuals need to adapt to that society in order to normalise their relations with the resident population over time. Individuals are in constant contact with the society due to the work done after migration, so it is thought that the work done has an important effect on the adaptation process. In this study, the effect of work on the post-migration adjustment process of individuals migrating from the first and second degree border neighbouring provinces of Van province was analysed using Poisson and Geographically Weighted Poisson Regression methods. The aim of the study is to determine the relationship between the contribution of work in the post-migration adjustment process and independent variables and to analyse which of the models used for the analysis gives stronger results. In the study, a face-to-face survey was conducted with 440 individuals and it was observed that the Geographically Weighted Poisson Regression method gave stronger results according to AIC, AICc and R^2 values. In addition, the effect and significance of the relationship between the dependent variable and independent variables according to provinces and districts are visualised and given with maps.

Using Spatial Modeling for Identifying Determinants Influencing Crash Estimate: Case Study of Hamedan, Iran

Road crashes are a major cause of fatalities worldwide, posing significant challenges for road-safety experts in selecting appropriate crash-frequency estimation models. This study introduces localized safety performance functions (C-SPFs), which explore the spatial variation of crash frequency and the spatial correlation between dependent variables. The exploratory spatial regression method is employed to identify optimal spatial associations. The study further predicts crashes using geographically weighted Poisson regression (GWPR) and generalized Poisson regression. Results indicate that C-SPFs offer greater accuracy than do models calibrated solely on annual average daily traffic. Moreover, the proposed model is especially relevant for jurisdictions facing higher heavy-vehicle traffic and frequent crashes. The development of C-SPFs and the use of GWPR provide valuable tools for policymakers and road-safety experts in enhancing crash-frequency estimation accuracy. Implementing these techniques can aid in prioritizing safety measures and countermeasures, especially in regions with significant heavy-vehicle traffic and crash occurrences. Additionally, the integration of spatial-analysis techniques and localized models can lead to more effective transportation planning and targeted road-safety interventions, ultimately contributing to reducing the burden of road crashes on a global scale.

Identifying spatial heterogeneity of COVID-19 transmission clusters and their built-environment features at the neighbourhood scale

The identification of high-risk areas for infectious disease transmission and its built-environment features are crucial for targeted surveillance and early prevention efforts. While previous research has explored the association between infectious disease incidence and urban built environment, the investigation of spatial heterogeneity of built-environment features in high-risk areas has been insufficient. This paper aims to address this gap by analysing the spatial heterogeneity of COVID-19 clusters in Shanghai at the neighbourhood scale and examining associated built-environment features. Using a spatiotemporal clustering algorithm, the study analysed 1395 reported cases in Shanghai from March 6 to March 17, 2022. Both global Poisson regression (GPR) and geographically weighted Poisson regression (GWPR) models were applied to examine the association between built-environment variables and the size of COVID-19 clusters. Our findings suggest that larger COVID-19 clusters emerging in the suburbs compared with the downtown and multiple built-environment features are significantly associated with this pattern. Specifically, neighbourhoods with a higher proportion of commercial, public service and industrial land, higher centrality of metro stations, and proximity to hospitals are positively associated with larger COVID-19 clusters, while neighbourhoods with higher land use mix and green/open spaces density are associated with smaller COVID-19 clusters. Moreover, we identified that metro stations with high centrality present the highest risk in the downtown, while commercial and public service places exhibit the highest risk in the suburbs. By highlighting the overlooked spatial heterogeneity of built-environment features for high-risk areas, this study aims to provide valuable guidance for public health departments in implementing place-based interventions to effectively prevent the spread of potential epidemics.

Handling Multicollinearity on Social Spatial Data Using Geographically Weighted Random Forest

Crime includes all kinds of harmful acts that violate the laws in force in Indonesia as well as social and religious norms. The crime total is the number of incidents reported to the police, obtained from public reports and events where the perpetrators were caught red-handed by the police. We can use the Poisson model to analyze the data, but the existence of spatial heterogeneity in the data makes the model less accurate. This research investigates the methods when there is spatial heterogeneity in the data by using Geographically weighted regression (GWR), Geographically Weighted Poisson Regression (GWPR) and Geographically Weighted Random Forest (GW-RF). We compare the GWR, GWPR, and GW-RF models for criminal cases in East Java in handling multicollinearity in the data. The results of this study indicate that the GW-RF model is better for modeling criminal cases with the smallest RMSE and MAPE values and an R-Square value close to 1. Based on the three most important variables in each location, they form six groups of regencies/cities in East Java, Indonesia. The variables vary between groups and the poverty severity index is not included in the three most important variables in all locations.

A linearization for stable and fast geographically weighted Poisson regression

Although geographically weighted Poisson regression (GWPR) is a popular regression for spatially indexed count data, its development is relatively limited compared to that found for linear geographically weighted regression (GWR), where many extensions (e.g. multiscale GWR, scalable GWR) have been proposed. The weak development of GWPR can be attributed to the computational cost and identification problem in the underpinning Poisson regression model. This study proposes linearized GWPR (L-GWPR) by introducing a log-linear approximation into the GWPR model to overcome these bottlenecks. Because the L-GWPR model is identical to the Gaussian GWR model, it is free from the identification problem, easily implemented, computationally efficient, and offers similar potential for extension. Specifically, L-GWPR does not require a double-loop algorithm, which makes GWPR slow for large samples. Furthermore, we extended L-GWPR by introducing ridge regularization to enhance its stability (regularized L-GWPR). The results of the Monte Carlo experiments confirmed that regularized L-GWPR estimates local coefficients accurately and computationally efficiently. Finally, we compared GWPR and regularized L-GWPR through a crime analysis in Tokyo.

Evaluating the spatial effects of environmental influencing factors on the frequency of urban crashes using the spatial Bayes method based on Euclidean distance and contiguity

Built environmental factors are one of the most important causes of urban crashes. Studies have shown that in addition to crash data, which have spatial heterogeneity, factors influencing crashes also have a spatial correlation. The main goal of this study is to evaluate the spatial effects of environmental factors on the frequency of crashes in Shiraz, Iran, at the TAZ level. In the first step of the study, using component analysis models, important environmental factors affecting the crash were identified, and composite indicators were produced as independent variables. In the second step, to control the effect of correlation and heterogeneity of model variables, spatial statistical models based on Euclidean distance such as geographically weighted Poisson regression (GWPR), geographically weighted negative binomial distribution (GWNBR), as well as Poisson and distribution models Negative binomial based on neighbor distance is used in spatial Bayes method with INLA approach. The study's results showed that models based on distance and contiguity to evaluate the spatial effects of crash data and the factors affecting it at the TAZ level have higher accuracy than geographically weighted regression models, as well as indicators of land use diversity and access to the system. The public transport produced in the first step effectively increases the frequency of crashes, and in TAZs where this index is high, there is a higher probability of a crash. The results of this study can be important for city managers and planners to improve urban safety measures, development planning, and future city measures.

The combined effects of high-speed rail and aviation on tourist arrivals in China

Most provincial tourism destinations have both high-speed rail (HSR) and aviation transportation systems; the combined effects of those two modes of transportation on tourist arrivals are likely to vary depending on location. This study examines panel data from 31 provinces in mainland China over a ten-year period (2010–2019) using the Geographically Weighted Poisson Regression (GWPR) model to quantify the combined effects of HSR and aviation on tourist arrivals and their spatiotemporal heterogeneity. The empirical results reveal that: first, HSR and aviation do have a comprehensive impact on tourist arrivals, but the development trends of the two are distinct, with the impact of HSR increasing while the impact of aviation is declining; second, the combined effects of HSR and aviation are clearly spatiotemporally heterogeneous, and their effects on both domestic and inbound tourists exhibit distinct states in various places at various times; third, the GWPR model is an effective tool for determining the combined effects of HSR and aviation on tourist arrivals as well as their regional heterogeneity. This paper not only broadens the research on the combined effects of HSR and aviation from a new perspective and effectively describes the temporal and spatial changes of the combined effect, but it also provides targeted development suggestions for different provincial regions.

Comparison of Global and Local Poisson Models for the Number of Recruitment Trees in Natural Forests

The recruitment of natural forests is the key to stand growth and regeneration. Constructing theoretical models for recruitment trees is crucial for accurately quantifying stand growth and yield. To this end, the objective was to use relevant Poisson models to study the spatial relationships between the number of recruitment trees (NRTs) and driving factors, such as topography, stand, and remote sensing factors. Taking the Northeast China Liangshui Nature Reserve as the study area and 127 ecological public welfare forest plots based on grid sampling as study data, we constructed global models (Poisson regression (PR) and linear mixed Poisson regression (LMPR)) and local models (geographically weighted Poisson regression (GWPR) and semiparametric GWPR (SGWPR)) to simulate the NRTs. The evaluation indicators were calculated to analyse four model fittings, predictive abilities, and spatial effects of residual analysis. The results show that local (GWPR and SGWPR) models have great advantages in all aspects. Compared with the GWPR model, the SGWPR model exhibited improved performance by considering whether coefficients have geographical variability for all independent variables. Therefore, the SGWPR model more accurately depicts the spatial distributions of NRTs than the other models.

Examining the Spatial Varying Effects of Sociodemographic Factors on Adult Cochlear Implantation Using Geographically Weighted Poisson Regression

To (i) demonstrate the utility of geographically weighted Poisson regression (GWPR) in describing geographical patterns of adult cochlear implant (CI) incidence in relation to sociodemographic factors in a publicly funded healthcare system, and (ii) compare Poisson regression and GWPR to fit the aforementioned relationship. Retrospective study of provincial CI Program database. Academic hospital. Adults 18 years or older who received a CI from 2020 to 2021. Cochlear implant. CI incidence based on income level, education attainment, age at implantation, and distance from center, and spatial autocorrelation across census metropolitan areas. Adult CI incidence varied spatially across Ontario (Moran's I = 0.04, p < 0.05). Poisson regression demonstrated positive associations between implantation and lower income level (coefficient = 0.0284, p < 0.05) and younger age (coefficient = 0.1075, p < 0.01), and a negative association with distance to CI center (coefficient = -0.0060, p < 0.01). Spatial autocorrelation was significant in Poisson model (Moran's I = 0.13, p < 0.05). GWPR accounted for spatial differences (Moran's I = 0.24, p < 0.690), and similar associations to Poisson were observed. GWPR further identified clusters of implantation in South Central census metropolitan areas with higher education attainment. Adult CI incidence demonstrated a nonstationary relationship between implantation and the studied sociodemographic factors. GWPR performed better than Poisson regression in accounting for these local spatial variations. These results support the development of targeted interventions to improve access and utilization to CIs in a publicly funded healthcare system.

Exploring Determinants of HIV/AIDS Self-Testing Uptake in South Africa Using Generalised Linear Poisson and Geographically Weighted Poisson Regression.

Increased HIV/AIDS testing is of paramount importance in controlling the HIV/AIDS pandemic and subsequently saving lives. Despite progress in HIV/AIDS testing programmes, most people are still reluctant to test and thus are still unaware of their status. Understanding the factors associated with uptake levels of HIV/AIDS self-testing requires knowledge of people's perceptions and attitudes, thus informing evidence-based decision making. Using the South African National HIV Prevalence, HIV Incidence, Behaviour and Communication Survey of 2017 (SABSSM V), this study assessed the efficacy of Generalised Linear Poisson Regression (GLPR) and Geographically Weighted Poisson Regression (GWPR) in modelling the spatial dependence and non-stationary relationships of HIV/AIDS self-testing uptake and covariates. The models were calibrated at the district level across South Africa. Results showed a slightly better performance of GWPR (pseudo R2 = 0.91 and AICc = 390) compared to GLPR (pseudo R2 = 0.88 and AICc = 2552). Estimates of local intercepts derived from GWPR exhibited differences in HIV/AIDS self-testing uptake. Overall, the output of this study displays interesting findings on the levels of spatial heterogeneity of factors associated with HIV/AIDS self-testing uptake across South Africa, which calls for district-specific policies to increase awareness of the need for HIV/AIDS self-testing.

Spatiotemporal pattern of COVID-19 mortality and its relationship with socioeconomic and environmental factors in England

This paper investigated the spatiotemporal pattern of COVID-19 mortality and its socioeconomic and environmental determinants in the first and second wave of the pandemic in England. The COVID-19 mortality rates for middle super output areas from March 2020 to April 2021 were used in the analysis. SaTScan was used in the analysis of spatiotemporal pattern of COVID-19 mortality and geographically weighted Poisson regression (GWPR) was used to investigate the association with socioeconomic and environmental factors. The results show that there was significant spatiotemporal variation in hotspots of COVID-19 deaths with the hotspots moving from regions where the COVID-19 outbreak initiated and then spread to other parts of the country. The GWPR analysis revealed that age composition, ethnic composition, deprivation, care home and pollution were all related to COVID-19 mortality. Althoughthe relationship varied over space the association with these factors was fairly consistent over the first and second wave.

Modeling the Number of Infant Mortality in East Lombok using Geographically Weighted Poisson Regression

Infant mortality is death that occurs at the age of 0 to 1 year. According to the Provincial Health Office, East Lombok is the district with the largest infant mortality rate in NTB. Several factors influence infant mortality: childbirth with medical assistance, low birth weight, health facilities, health workers, and exclusive breastfeeding. These factors have a spatial influence because each region has different geographical, socio-cultural, and economic conditions. Therefore, the method that can be used is GWPR because it can model data with the response variable with a Poisson distribution and pay attention to location or spatial aspects. This study aims to determine the infant mortality model in East Lombok using Geographically Weighted Poisson Regression (GWPR) and to determine the factors that significantly influence the number of infant deaths in East Lombok. Based on the research conducted showed that low birth weight is the only factor that significantly affected infant mortality in 8 sub-districts, including Keruak, Sakra, West Sakra, East Sakra, Terara, Sukamulia, Selong, and Labuhan Haji. The model obtained gives a good estimator, with an R^2 value of 76,44%.

Spatiotemporal heterogeneity of SARS-CoV-2 diffusion at the city level using geographically weighted Poisson regression model: The case of Bologna, Italy.

This paper aimed to analyse the spatio-temporal patterns of the diffusion of SARS-CoV-2, the virus causing coronavirus 2019 (COVID-19, in the city of Bologna, the capital and largest city of the Emilia-Romagna Region in northern Italy. The study took place from February 1st, 2020 to November 20th, 2021 and accounted for space, sociodemographic characteristics and health conditions of the resident population. A second goal was to derive a model for the level of risk of being infected by SARS-CoV-2 and to identify and measure the place-specific factors associated with the disease and its determinants. Spatial heterogeneity was tested by comparing global Poisson regression (GPR) and local geographically weighted Poisson regression (GWPR) models. The key findings were that different city areas were impacted differently during the first three epidemic waves. The area-to-area influence was estimated to exert its effect over an area with 4.7 km radius. Spatio-temporal heterogeneity patterns were found to be independent of the sociodemographic and the clinical characteristics of the resident population. Significant single-individual risk factors for detected SARS-CoV-2 infection cases were old age, hypertension, diabetes and co-morbidities. More specifically, in the global model, the average SARS-CoV-2 infection rate decreased 0.93-fold in the 21-65 years age group compared to the >65 years age group, whereas hypertension, diabetes, and any other co-morbidities (present vs absent), increased 1.28-, 1.39- and 1.15-fold, respectively. The local GWPR model had a better fit better than GPR. Due to the global geographical distribution of the pandemic, local estimates are essential for mitigating or strengthening security measures.

Segment-Level Spatial Heterogeneity of Arterial Crash Frequency Using Locally Weighted Generalized Linear Models

Decades of literature in traffic crash modeling show the popularity of generalized linear models (GLMs). However, because of the failure to accommodate spatial heterogeneity, parameters estimated with these models are inconsistent and inefficient. In light of that, this study aims to investigate the spatial heterogeneity of crashes aggregated at roadway segment levels using geographically weighted Poisson regression (GWPR) and two variants of the geographically weighted negative binomial regression (GWNBR) model. The results indicate: (i) the GWNBR model with global dispersion parameter outperforms conventional Poisson, GWPR, and negative binomial (NB) models; (ii) the performance of the GWNBR model further enhances as the dispersion parameter becomes spatially non-stationary; (iii) tests of spatial heterogeneity and autocorrelation reveal the existence of non-stationarity and less than 1% likelihood of randomness; and (iv) median of parameter estimates reveal a positive association between crashes and posted speed limit, number of lanes, number of three-leg intersections, number of access points, and vehicle miles traveled (VMT). The study concludes that, when spatial heterogeneity is evident, conventional GLMs should be avoided to circumvent the futile estimation of parameters on the arterial segments. The findings are expected to contribute to the small pool of literature on spatial non-stationarity of parameters in segment-based aggregation, identification and selection of segments with similar influence factors in corridor-level studies, and spatial analysis of arterial crashes.