Hierarchical Spatial Model Research Articles

Exploring green gentrification in 28 global North cities: the role of urban parks and other types of greenspaces

Although cities globally are increasingly mobilizing re-naturing projects to address diverse urban socio-environmental and health challenges, there is mounting evidence that these interventions may also be linked to the phenomenon known as green gentrification. However, to date the empirical evidence on the relationship between greenspaces and gentrification regarding associations with different greenspace types remains scarce. This study focused on 28 mid-sized cities in North America and Western Europe. We assessed improved access to different types of greenspace (i.e. total area of parks, gardens, nature preserves, recreational areas or greenways [i] added before the 2000s or [ii] added before the 2010s) and gentrification processes (including [i] gentrification for the 2000s; [ii] gentrification for the 2010s; [iii] gentrification throughout the decades of the 2000s and 2010s) in each small geographical unit of each city. To estimate the associations, we developed a Bayesian hierarchical spatial model for each city and gentrification time period (i.e. a maximum of three models per city). More than half of our models showed that parks—together with other factors such as proximity to the city center—are positively associated with gentrification processes, particularly in the US context, except in historically Black disinvested postindustrial cities with lots of vacant land. We also find than in half of our models newly designated nature preserves are negatively associated with gentrification processes, particularly when considering gentrification throughout the 2000s and the 2010s and in the US. Meanwhile, for new gardens, recreational spaces and greenways, our research shows mixed results (some positive, some negative and some no effect associations). Considering the environmental and health benefits of urban re-naturing projects, cities should keep investing in improving park access while simultaneously implementing anti-displacement and inclusive green policies.

Community-level characteristics of COVID-19 vaccine hesitancy in England: A nationwide cross-sectional study

One year after the start of the COVID-19 vaccination programme in England, more than 43 million people older than 12 years old had received at least a first dose. Nevertheless, geographical differences persist, and vaccine hesitancy is still a major public health concern; understanding its determinants is crucial to managing the COVID-19 pandemic and preparing for future ones. In this cross-sectional population-based study we used cumulative data on the first dose of vaccine received by 01-01-2022 at Middle Super Output Area level in England. We used Bayesian hierarchical spatial models and investigated if the geographical differences in vaccination uptake can be explained by a range of community-level characteristics covering socio-demographics, political view, COVID-19 health risk awareness and targeting of high risk groups and accessibility. Deprivation is the covariate most strongly associated with vaccine uptake (Odds Ratio 0.55, 95%CI 0.54-0.57; most versus least deprived areas). The most ethnically diverse areas have a 38% (95%CI 36-40%) lower odds of vaccine uptake compared with those least diverse. Areas with the highest proportion of population between 12 and 24 years old had lower odds of vaccination (0.87, 95%CI 0.85-0.89). Finally increase in vaccine accessibility is associated with COVID-19 vaccine coverage (OR 1.07, 95%CI 1.03-1.12). Our results suggest that one year after the start of the vaccination programme, there is still evidence of inequalities in uptake, affecting particularly minorities and marginalised groups. Strategies including prioritising active outreach across communities and removing practical barriers and factors that make vaccines less accessible are needed to level up the differences.

Delineating urban functional zones using mobile phone data: A case study of cross-boundary integration in Shenzhen-Dongguan-Huizhou area

As cities continuously expand and with the emergence of mega-city regions, the urban functional zones (UFZs) have spread beyond their original administrative boundaries. An accurate and updated delineation of the UFZs is crucial for assessing the functional integration between cities within a mega-city region. Mobility data provides a chance to depict the UFZs from actual human activities at a finer spatial scale. Existing studies mostly adopted network-based approaches relying on the topological relationship but ignoring spatial factors, causing the lack of sensitivity in detecting the cross-cities integration of the functional region. This research proposed a novel regionalisation algorithm that redraws non-overlap boundaries of urban functional zones based on the commuting origin-destination matrix, representing the spatial interactions within cities and cross-cities. In particular, functional zones are drawn by searching for the best partition with the best goodness of fitting in the hierarchical spatial interaction model. The algorithm was applied to a case study of a mega-city region, Shenzhen-Dongguan-Huizhou (SDH) area in China using mobile phone signalling data. By adopting two different settings, this model evaluated the current status and predict the future trend of urban integration respectively. The results show the current boundary of UFZs in the SDH area almost coincides with administrative boundaries. Meanwhile, the results of long-term predictions might be utilised by policymakers to give more attention to the areas near the Dongguan-Huizhou boundary to promote industry cooperation and avoid mismatches between the functional and administrative regions.

Prediction and Surveillance Sampling Assessment in Plant Nurseries and Fields

In this paper, we propose a structured additive regression (STAR) model for modeling the occurrence of a disease in fields or nurseries. The methodological approach involves a Gaussian field (GF) affected by a spatial process represented by an approximation to a Gaussian Markov random field (GMRF). This modeling allows the building of maps with prediction probabilities regarding the presence of a disease in plants using Bayesian kriging. The advantage of this modeling is its computational benefit when compared with known spatial hierarchical models and with the Bayesian inference based on Markov chain Monte Carlo (MCMC) methods. Inference through the use of the integrated nested Laplace approximation (INLA) with the stochastic partial differential equation (SPDE) approach facilitates the handling of large datasets in excellent computation times. Our approach allows the evaluation of different sampling strategies, from which we obtain inferences and prediction maps with similar behaviour to those obtained when we consider all subjects in the study population. The analysis of the different sampling strategies allows us to recognize the relevance of spatial components in the studied phenomenon. We demonstrate how Bayesian kriging can incorporate sources of uncertainty associated with the prediction parameters, which leads to more realistic and accurate estimation of the uncertainty. We illustrate the methodology with samplings of Citrus macrophylla affected by the tristeza virus (CTV) grown in a nursery.

High-resolution Bayesian mapping of landslide hazard with unobserved trigger event

Statistical models for landslide hazard enable mapping of risk factors and landslide occurrence intensity by using geomorphological covariates available at high spatial resolution. However, the spatial distribution of the triggering event (e.g., precipitation or earthquakes) is often not directly observed. In this paper we develop Bayesian spatial hierarchical models for point patterns of landslide occurrences using different types of log-Gaussian Cox processes. Starting from a competitive baseline model that captures the unobserved precipitation trigger through a spatial random effect at slope unit resolution, we explore novel complex model structures that take clusters of events arising at small spatial scales into account as well as nonlinear or spatially-varying covariate effects. For a 2009 event of around 5000 precipitation-triggered landslides in Sicily, Italy, we show how to fit our proposed models efficiently, using the integrated nested Laplace approximation (INLA), and rigorously compare the performance of our models both from a statistical and applied perspective. In this context we argue that model comparison should not be based on a single criterion and that different models of various complexity may provide insights into complementary aspects of the same applied problem. In our application our models are found to have mostly the same spatial predictive performance, implying that key to successful prediction is the inclusion of a slope-unit resolved random effect capturing the precipitation trigger. Interestingly, a parsimonious formulation of space-varying slope effects reflects a physical interpretation of the precipitation trigger: in subareas with weak trigger, the slope steepness is shown to be mostly irrelevant.

Associations Between the Physical Availability of Alcohol and Alcohol Use: Regional Variation Across 15 Major Cities in Ontario, Canada.

Rates of alcohol use and alcohol-related harms increase with greater alcohol availability. However, regional differences in sociodemographic characteristics and built environment may affect this association. This study evaluated the association between off-premise alcohol availability and alcohol use in Ontario, Canada, and the degree to which this association varies between cities. This was a cross-sectional spatial analysis of urban neighborhoods in Ontario, Canada (n = 11,742). The primary exposure was off-premise alcohol availability, based on the drive time from a neighborhood to the closest off-premise outlets. The primary outcome was the neighborhood-level prevalence of high-volume alcohol use (>2 drinks/day [males], >1 drink/ day [females]) based on survey data from 2000 to 2014. The association between availability and use was assessed using Bayesian hierarchical spatial models to account for spatial autocorrelation. There was an overall positive association between alcohol availability and high-volume alcohol use (male coefficient estimate (β) = 0.19, 95% credible interval [CI] [0.16, 0.22]; female β = 0.17, 95% CI [0.13, 0.21]). However, the association was eliminated in models that allowed for this association to vary between cities via an interaction term (male β = -0.04, 95% CI [-0.26, 0.19]; female β = -0.04, 95% CI [-0.34, 0.26]). This was explained by variability in the association between cities, where some cities demonstrated a positive association between availability and use and others demonstrated a negative association. Although there is a province-wide positive association between off-premise alcohol availability and high-volume alcohol use, there is substantial regional variation in this association that may affect the local effectiveness of alcohol regulation policies.

Age- and Sex-Specific Association Between Vegetation Cover and Mental Health Disorders: Bayesian Spatial Study.

BackgroundDespite growing evidence that reduced vegetation cover could be a putative risk factor for mental health disorders, the age- and the sex-specific association between vegetation and mental health disorder cases in urban areas is poorly understood. However, with rapid urbanization across the globe, there is an urgent need to study this association and understand the potential impact of vegetation loss on the mental well-being of urban residents.ObjectiveThis study aims to analyze the spatial association between vegetation cover and the age- and sex-stratified mental health disorder cases in the neighborhoods of Toronto, Canada.MethodsWe used remote sensing to detect urban vegetation and Bayesian spatial hierarchical modeling to analyze the relationship between vegetation cover and mental health disorder cases. Specifically, an Enhanced Vegetation Index was used to detect urban vegetation, and Bayesian Poisson lognormal models were implemented to study the association between vegetation and mental health disorder cases of males and females in the 0-19, 20-44, 45-64, and ≥65 years age groups, after controlling for marginalization and unmeasured (latent) spatial and nonspatial covariates at the neighborhood level.ResultsThe results suggest that even after adjusting for marginalization, there were significant age- and sex-specific effects of vegetation on the prevalence of mental health disorders in Toronto. Mental health disorders were negatively associated with the vegetation cover for males aged 0-19 years (−7.009; 95% CI −13.130 to −0.980) and for both males (−4.544; 95% CI −8.224 to −0.895) and females (−3.513; 95% CI −6.289 to −0.681) aged 20-44 years. However, for older adults in the 45-64 and ≥65 years age groups, only the marginalization covariates were significantly associated with mental health disorder cases. In addition, a substantial influence of the unmeasured (latent) and spatially structured covariates was detected in each model (relative contributions>0.7), suggesting that the variations in area-specific relative risk were mainly spatial in nature.ConclusionsAs significant and negative associations between vegetation and mental health disorder cases were found for young males and females, investments in urban greenery can help reduce the future burden of mental health disorders in Canada. The findings highlight the urgent need to understand the age-sex dynamics of the interaction between surrounding vegetation and urban dwellers and its subsequent impact on mental well-being.

Infrastructure and income inequality: An application to the Brazilian case using hierarchical spatial autoregressive models

AbstractMany scholars have highlighted the role of infrastructure in reducing income inequality. Developing economies present immense regional and income discrepancies, which are correlated with unequally distributed infrastructure in territorial and population terms. In this paper, we assess the effects of infrastructure supply on income inequality and verify whether these effects vary according to the infrastructure sector and its degree of quality and access in Brazil. The analysis is based on spatial hierarchical models. Results allow us to say that infrastructure correlates negatively with income inequality. Hence, policies aimed at improving infrastructure quality and expanding access are crucial for reducing income concentration.

On Hierarchical Bayesian Spatial Small Area Model for Binary Data under Spatial Misalignment

Small area models have become popular methods for producing reliable estimates for sub-populations (small geographic areas in this study). Small area modeling may be carried out via model-assisted approaches within the model-based approaches or design-based paradigm. When there are medium or large samples, a model-assisted approach may be reliable. However, when data are scarce, a model-based technique may be required. Model-based Bayesian analysis is popular for its ability to combine information from several sources as well as taking account uncertainties in the analysis and spatial prediction of spatial data. Nevertheless, things become more complex when the geographic boundaries of interest are misaligned. Some authors have addressed the problem of misalignment under hierarchical Bayesian approach. In this study, we developed non-trivial extension of existing hierarchical Bayesian model for a binary outcome variable under spatial misalignment with three contributions. First, the model uses unit-level survey data and area-level auxiliary data to predict the posterior mean proportion spatially at the second geographic area level. Second, the linking model is changed to logit-normal model in the proposed model. Lastly, the mean process was considered to overcome the multicollinearity between the true predictors and the spatial random effect. Sensitivity analysis was also done via simulation.

Contagion effects of UK small business failures: A spatial hierarchical autoregressive model for binary data

This article focuses on modelling the contagion effects - both between and within groups - on small business failures in London. Small business clusters can be defined based on different companies’ characteristics, for example, economic sector or geographical location. These aspects are usually included as fixed effects to predict the defaults of small- and medium-sized enterprises (SMEs). However, this approach however ignores the interactions between the company groups and only captures the heterogeneity across the clusters. To include both contagion effects between and within groups, a Bayesian spatial hierarchical model for binary data is proposed and applied to a dataset of SMEs located in London in 2016. The empirical analysis shows that the contagion component at the lower level, based on the geographical location, is not significant if the industry clustering is ignored. However, it becomes significant if the industry group effect is included, and also the upper-level interdependence also becomes significant. Finally, the suggested model improves the predictive accuracy and the expected shortfall estimate compared to standard scoring models.

Stream size, temperature, and density explain body sizes of freshwater salmonids across a range of climate conditions

Climate change and anthropogenic activities are altering the body sizes of fishes, yet our understanding of factors influencing body size for many taxa remains incomplete. We evaluated the relationships between climate, environmental, and landscape attributes and the body size of different taxa of freshwater trout (Salmonidae) in the USA. Hierarchical spatial modeling across a gradient of habitats (5221 sites) illustrated the importance of watershed effects, which explained 17%–45% of the of the variation in body size across taxa. Stream size had a strong, positive relationship with body size, yet there was approximately tenfold difference in the strength of the relationship across taxa. Trout body size consistently declined with increasing density across taxa. Despite reliance on cold water, we found positive relationships between summer stream temperature and trout body size across most taxa. Our results highlight how providing trout access to larger, productive rivers for the expression of growth and life-history variation would promote body size diversity within and across populations.

A Nearest-Neighbour Gaussian Process Spatial Factor Model for Censored, Multi-Depth Geochemical Data

Abstract We investigate the relationships between local environmental variables and the geochemical composition of the Earth in a region spanning over 26,000 km2 in the lower South Island of New Zealand. Part of the Southland–South Otago geochemical baseline survey—a pilot study pre-empting roll-out across the country—the data comprise the measurements of 59 chemical trace elements, each at two depth prescriptions, at several hundred spatial sites. We demonstrate construction of a hierarchical spatial factor model that captures inter-depth dependency; handles imputation of left-censored readings in a statistically principled manner; and exploits sparse approximations to Gaussian processes to deliver inference. The voluminous results provide a novel impression of the underlying processes and are presented graphically via simple web-based applications. These both confirm existing knowledge and provide a basis from which new research hypotheses in geochemistry might be formed.

Estimating confidence intervals for spatial hierarchical mixed-effects models with post-stratification

Analyzing population representative datasets for local level estimation and prediction purposes is important for monitoring public health, however, there are many statistical challenges associated with such analyses. Small area estimation (SAE) with post-stratified hierarchical mixed-effects models is a popular method for analysis. Post-stratification is a method that creates area-level predictions from a model fitting using sub-area-level covariates by incorporating auxiliary information (i.e., census data). While the post-stratification is an intuitive approach, the predictive benefits of post-stratification over standard methods with hierarchical mixed-effects models remain unclear. Another challenge for analyzing this type of data is the incorporation of sampling weights, as common data sources utilize complex sampling designs with uneven sampling probabilities. In addition, estimating the mean squared prediction error (MSPE) can be difficult via asymptotic theory due to the complex sampling designs and post-stratification process. Bootstrap methods can be an alternative, however there are many bootstrapping methods to choose from and their properties in realistic scenarios are unclear. In this paper, we compared the predictive ability of post-stratified and non-post-stratified estimators and evaluate the performance of various bootstrapping methods in estimating the MSPE with simulated data. Further, we compare the results using a population-based survey used to estimate the county-level prevalence of smoking in the state of South Carolina.

Measuring the process of urban gentrification: A composite measure of the gentrification process in Toronto

Gentrification is being experienced by cities around the world. Its drivers and characteristic features are complex and diverse, ranging from the displacement of low-income households to the redevelopment of commercial districts. This paper combines multiple data sources to explore the coevolution of gentrification in the residential market and developments in non-residential sectors in the Canadian city of Toronto. Analysis starts from a max-p-regions clustering based on a composite measure of the gentrification process, which includes measures of household income, educational attainment, building permits, and the composition of establishments. The variables that describe each region are examined for patterns of gentrification. We then develop a Bayesian hierarchical spatial (BHS) model to describe the change in residential property prices over the five-year period between 2011 and 2016. The results show that establishment entropy and composition are found to influence residential property prices, with significant spatial variation in its effect. The combination of endogenous region definitions and hierarchical modeling is strongly supported by model results.

Geographic Variation in HIV Testing Among Transgender and Nonbinary Adults in the United States.

Transgender and nonbinary (TNB) populations are disproportionately affected by HIV and few local health departments or HIV surveillance systems collect/report data on TNB identities. Our objective was to estimate the prevalence of HIV testing among TNB adults by US county and state, with a focus on the Ending the HIV Epidemic (EHE) geographies. We applied a Bayesian hierarchical spatial small area estimation model to data from the 2015 US Transgender Survey, a large national cross-sectional Internet-based survey. We estimated the county- and state-level proportion of TNB adults who ever tested or tested for HIV in the last year by gender identity, race/ethnicity, and age. Our analysis included 26,100 TNB participants with valid zip codes who resided in 1688 counties (54% of all 3141 counties that cover 92% of the US population). The median county-level proportion of TNB adults who ever tested for HIV was 44% (range 10%-80%) and who tested in the last year was 17% (range 4%-44%). Within most counties, testing was highest among transgender women, black respondents, and people aged ≥25 years. HIV testing was lowest among nonbinary people and young adults aged <25 years. The proportion of TNB adults who tested within the last year was very low in most EHE counties and in all 7 rural states. HIV testing among TNB adults is likely below national recommendations in the majority of EHE geographies. Geographic variation in HIV testing patterns among TNB adults indicates that testing strategies need to be tailored to local settings.

Intra-urban variation in tuberculosis and community socioeconomic deprivation in Lisbon metropolitan area: a Bayesian approach

BackgroundMultidrug resistant tuberculosis (MDR-TB) is a recognized threat to global efforts to TB control and remains a priority of the National Tuberculosis Programs. Additionally, social determinants and socioeconomic deprivation have since long been associated with worse health and perceived as important risk factors for TB. This study aimed to analyze the spatial distribution of non-MDR-TB and MDR-TB across parishes of the Lisbon metropolitan area of Portugal and to estimate the association between non-MDR-TB and MDR-TB and socioeconomic deprivation.MethodsIn this study, we used hierarchical Bayesian spatial models to analyze the spatial distribution of notification of non-MDR-TB and MDR-TB cases for the period from 2000 to 2016 across 127 parishes of the seven municipalities of the Lisbon metropolitan area (Almada, Amadora, Lisboa, Loures, Odivelas, Oeiras, Sintra), using the Portuguese TB Surveillance System (SVIG-TB). In order to characterise the populations, we used the European Deprivation Index for Portugal (EDI-PT) as an indicator of poverty and estimated the association between non-MDR-TB and MDR-TB and socioeconomic deprivation.ResultsThe notification rates per 10,000 population of non-MDR TB ranged from 18.95 to 217.49 notifications and that of MDR TB ranged from 0.83 to 3.70. We identified 54 high-risk areas for non-MDR-TB and 13 high-risk areas for MDR-TB. Parishes in the third [relative risk (RR) = 1.281, 95% credible interval (CrI): 1.021–1.606], fourth (RR = 1.786, 95% CrI: 1.420–2.241) and fifth (RR = 1.935, 95% CrI: 1.536–2.438) quintile of socioeconomic deprivation presented higher non-MDR-TB notifications rates. Parishes in the fourth (RR = 2.246, 95% CrI: 1.374–3.684) and fifth (RR = 1.828, 95% CrI: 1.049–3.155) quintile of socioeconomic deprivation also presented higher MDR-TB notifications rates.ConclusionsWe demonstrated significant heterogeneity in the spatial distribution of both non-MDR-TB and MDR-TB at the parish level and we found that socioeconomically disadvantaged parishes are disproportionally affected by both non-MDR-TB and MDR-TB. Our findings suggest that the emergence of MDR-TB and transmission are specific from each location and often different from the non-MDR-TB settings. We identified priority areas for intervention for a more efficient plan of control and prevention of non-MDR-TB and MDR-TB.Graphical

Targeting the search of African swine fever-infected wild boar carcasses: A tool for early detection.

This study analyses the temporal and spatial distribution of found dead African swine fever (ASF)-positive wild boar carcasses from 2017 to January 2021 in affected European countries: Bulgaria, Estonia, Germany, Hungary, Latvia, Lithuania, Romania, Poland, Serbia and Slovakia. During this period, a total of 21,785 cases were confirmed in 19,071 unique locations. The temporal analysis of aggregated cases per month evidenced that most countries located in southern latitudes showed a higher number of cases between January and April, whereas in northern latitudes there was no clear temporal pattern. The space-time K-function evidenced a space-time clustering in the ASF-positive wild boar carcasses, which was most prominent within distances of 2 km and within 1week. A Bayesian hierarchical spatial model was calibrated to evaluate the association between the probability of finding ASF-positive wild boar carcasses and landscape factors (i.e. the presence of a path and paved road), land use and wild boar abundance. Results showed the highest likelihood of finding ASF-positive wild boar carcasses in areas of transition between woodland and shrub, green urban areas and mixed forests. The presence of a path and a higher abundance of wild boar also increased slightly the odds of finding an ASF-positive dead wild boar. In summary, this paper aims to provide recommendations to design a search strategy to find ASF-infected wild boar carcasses, which is a crucial activity in the management of the disease, not just for surveillance purposes (i.e. the early detection of an introduction and the regular monitoring to understand the epidemiology and dynamics), but also for control, namely the disposal of infected carcasses as a virus source.

A Bayesian Hierarchical Spatial Model to Correct for Misreporting in Count Data: Application to State-Level COVID-19 Data in the United States.

The COVID-19 pandemic that began at the end of 2019 has caused hundreds of millions of infections and millions of deaths worldwide. COVID-19 posed a threat to human health and profoundly impacted the global economy and people’s lifestyles. The United States is one of the countries severely affected by the disease. Evidence shows that the spread of COVID-19 was significantly underestimated in the early stages, which prevented governments from adopting effective interventions promptly to curb the spread of the disease. This paper adopts a Bayesian hierarchical model to study the under-reporting of COVID-19 at the state level in the United States as of the end of April 2020. The model examines the effects of different covariates on the under-reporting and accurate incidence rates and considers spatial dependency. In addition to under-reporting (false negatives), we also explore the impact of over-reporting (false positives). Adjusting for misclassification requires adding additional parameters that are not directly identified by the observed data. Informative priors are required. We discuss prior elicitation and include R functions that convert expert information into the appropriate prior distribution.

Interpolation of Precipitation Extremes on a Large Domain Toward IDF Curve Construction at Unmonitored Locations

An intensity–duration–frequency (IDF) curve describes the relationship between rainfall intensity and duration for a given return period and location. Such curves are obtained through frequency analysis of rainfall data and commonly used in infrastructure design, flood protection, water management, and urban drainage systems. However, they are typically available only in sparse locations. Data for other sites must be interpolated as the need arises. This paper describes how extreme precipitation of several durations can be interpolated to compute IDF curves on a large, sparse domain. In the absence of local data, a reconstruction of the historical meteorology is used as a covariate for interpolating extreme precipitation characteristics. This covariate is included in a hierarchical Bayesian spatial model for extreme precipitations. This model is especially well suited for a covariate gridded structure, thereby enabling fast and precise computations. As an illustration, the methodology is used to construct IDF curves over Eastern Canada. An extensive cross-validation study shows that at locations where data are available, the proposed method generally improves on the current practice of Environment and Climate Change Canada which relies on a moment-based fit of the Gumbel extreme-value distribution.

Bayesian nonparametric multivariate spatial mixture mixed effects models with application to American Community Survey special tabulations

Leveraging multivariate spatial dependence to improve the precision of estimates using American Community Survey data and other sample survey data has been a topic of recent interest among data users and federal statistical agencies. One strategy is to use a multivariate spatial mixed effects model with a Gaussian observation model and latent Gaussian process model. In practice, this works well for a wide range of tabulations. Nevertheless, in situations in which the data exhibit heterogeneity within or across geographies, and/or there is sparsity in the data, the Gaussian assumptions may be problematic and lead to underperformance. To remedy these situations, we propose a multivariate hierarchical Bayesian nonparametric mixed effects spatial mixture model to increase model flexibility. The number of clusters is chosen automatically in a data-driven manner. The effectiveness of our approach is demonstrated through a simulation study and motivating application of special tabulations for American Community Survey data.