Spatial autocorrelation and data uncertainty in the American Community Survey: a critique

ABSTRACT

We argue that the use of American Community Survey (ACS) data in spatial autocorrelation statistics without considering error margins is critically problematic. Public health and geographical research has been slow to recognize high data uncertainty of ACS estimates, even though ACS data are widely accepted data sources in neighborhood health studies and health policies. Detecting spatial autocorrelation patterns of health indicators on ACS data can be distorted to the point that scholars may have difficulty in perceiving the true pattern. We examine the statistical properties of spatial autocorrelation statistics of areal incidence rates based on ACS data. In a case study of teen birth rates in Mecklenburg County, North Carolina, in 2010, Global and Local Moran’s I statistics estimated on 5-year ACS estimates (2006–2010) are compared to ground truth rate estimates on actual counts of births certificate records and decennial-census data (2010). Detected spatial autocorrelation patterns are found to be significantly different between the two data sources so that actual spatial structures are misrepresented. We warn of the possibility of misjudgment of the reality and of policy failure and argue for new spatially explicit methods that mitigate the biasedness of statistical estimations imposed by the uncertainty of ACS data.     

Patterns of mortality in Freetown,Sierra Leone

A demographic and spatial analysis of mortality in Freetown, Sierra Leone, is presented, with emphasis on data for the census years 1963 and 1974. Consideration is first given to the availability of data and their quality. The authors then examine geographic variations in mortality within the city, age- and sex-specific death rates, the major causes of death, and death certification.     

High resolution dasymetric model of U.S demographics with application to spatial distribution of racial diversity

Population and demographic data at high spatial resolution is a valuable resource for supporting planning and management decisions as well as an important input to socio-economic academic studies. Dasymetric modeling has been a standard technique to disaggregate census-aggregated units into raster-based data of higher spatial resolution. Although utility of dasymetric mapping has been demonstrated on local and regional scales, few high resolution large-scale models exist due to their high computational cost. In particular, no publicly available high resolution dasymetric model of population distribution over the entire United States is presently available. In this paper we introduce a 3″ (∼90 m) resolution dasymetric model of demographics over the entire conterminous United States. The major innovation is to disaggregate already existing 30″ (∼1 km) and 7.5″ (∼250 m) SEDAC (Socioeconomic Data and Applications Center) Census 2000 grids instead of the original census block-level data. National Land Cover Dataset (NLCD) 2001 is used as ancillary information. This allows for rapid development of a U.S.-wide model for distribution of population and sixteen other demographic variables. The new model is demonstrated to markedly improve spatial accuracy of SEDAC model. To underscore importance of high spatial resolution demographic information other than total population count we demonstrate how maps of several population characteristics can be fused into a “product” map that illustrates complex social issues. Specifically, we introduce a “diversity” categorical map that informs (at nominal 3″ resolution) about spatial distribution of racial diversity, dominant race, and population density simultaneously. Diversity map is compared to a similar map based on census tracts. High resolution raster map allows study of race-diversity phenomenon on smaller scale, and, outside of major metropolitan areas, revels existence of patterns that cannot be deduced from a tract-based map. The new high resolution population and diversity maps can be explored online using our GeoWeb application DataEye available at http://sil.uc.edu/. Both datasets can be also downloaded from the same website.     

Incorporating GIS Building Data and Census Housing Statistics for Sub-Block-Level Population Estimation

This article presents a deterministic model for sub-block-level population estimation based on the total building volumes derived from geographic information system (GIS) building data and three census block-level housing statistics. To assess the model, we generated artificial blocks by aggregating census block areas and calculating the respective housing statistics. We then applied the model to estimate populations for sub-artificial-block areas and assessed the estimates with census populations of the areas. Our analyses indicate that the average percent error of population estimation for sub-artificial-block areas is comparable to those for sub-census-block areas of the same size relative to associated blocks. The smaller the sub-block-level areas, the higher the population estimation errors. For example, the average percent error for residential areas is approximately 0.11 percent for 100 percent block areas and 35 percent for 5 percent block areas.     

Unemployment in Non-Metropolitan Australia: integrating geography,social and individual contexts

Despite a significant period of strong economic and jobs growth nationally, there is well-established evidence in Australia that the proceeds of this growth have not been shared equally, either between places or between individuals. Empirically, it is well known that particular socioeconomic groups have a higher risk of unemployment and it has become equally well established that there are particular geographic patterns of labour market disadvantage that suggest that local geographic context is also important. What is not well understood are the ways in which phenomena at the geographic level are associated with individual-level characteristics and other social contexts in ways that negatively impact on a range of social outcomes, including unemployment. The present paper specifically addresses this issue by using a multi-scalar approach and using survey data from the Housing, Income and Labour Force Dynamics Australia (HILDA) survey and aggregate level census data to model unemployment risk. The paper argues that to better understand unemployment and to add to sound policy development, approaches that incorporate a variety of contexts, including the impact of local geographies, are important.     

The Council at Work

Abstract

Through analysis of census data, these lessons cover geographic concepts dealing with migration and population change in the United States. Students discuss the historical push and pull factors of immigration to the United States. By focusing on the recent influx of Hispanic immigrants, students look at the geographic concepts of assimilation, discrimination, and time-decay. Students also create graphs and maps to examine the recent increase in the United States Hispanic population and geographic patterns of Hispanic settlement.     

1951年以来奥地利人口变化的空间特征

根据1951-2001年奥地利人口调查统计数据,分析了近50年来奥地利人口变化的时空间规律,以及人口变化中的增长区和下降区的区域配置等特征,并建立人口变化模式。研究表明：20世纪后半叶奥地利经历了其他欧盟国所表现出的城市化、城市郊区化和人口老龄化的各个过程。特别是城市老龄化表现更为突出,据预测推算,到2035年奥地利将会有1/3的人口超过60岁。在分析过程中,不仅应用了行政区划的区域方法,而且也依据土地利用的经济性质,在“土地经济生产小区”的基底上,对人口的变化规律进行了更为深入的透视。研究结果显示：奥地利西部阿尔卑斯山地中的城市区域人口增长幅度最大。沿着捷克和斯洛伐克边界伸展的北部和东北部区域是最严重的人口缺失地区。而奥地利阿尔卑斯山地东部的一些早期产业都市及其郊区,乃至更远一些的南部和东南部的老工业区域,目前由于普遍存在着经济的困境,也直接导致了这些地区人口的下降。但奥地利未来的总人口趋势将持续稳定。     

Estimating the Accuracy of Geographic Variation in Indigenous Population Counts

The ABS 2006 Post-Enumeration Survey was extended to include a sample of localities from the whole of Australia, thereby providing an estimate of census net undercount reflective of the enumeration in remote Indigenous settlements for the first time. The results revealed substantial undercounting of the Indigenous population in certain jurisdictions. At the same time, census counts in many locations were substantially higher than demographic factors could account for. The analytical and policy issues that arise from this revolve around a simple question: what credence can we give to observed spatial patterns of Indigenous population change? This paper seeks to provide an answer by establishing the spatial relationship between population change and net migration at the small area level. This reveals the detailed geography of census undercount and ‘overcount’ with the former common in remote areas and the latter most evident in regional towns and cities. The findings raise important policy issues about the proper interpretation of Indigenous population change data and the nature of estimates used to determine fiscal resourcing for Indigenous policies and programs.     

The residential segregation of San Antonio,Texas in 1910: an analysis of ethno-racial and occupational spatial patterns with the colocation quotient

The segregation of cities can be traced to a time when the compartmentalization of space and people was based on factors other than race. In segregation research, one of the limiting factors has always been the geographic scale of the data, and the limited knowledge that exists of segregation patterns when the household is the unit of analysis. Historical census data provides the opportunity to analyze the disaggregated information, and this paper does so with San Antonio during 1910. A spatial analysis of residential segregation based on race, ethnicity, and occupations is carried out with the colocation quotient to map and measure the attraction of residents. Results reveal the presence of residential segregation patterns on different sectors of the city based on households’ ethno-racial and occupational attributes; therefore, providing evidence of the existence of residential segregation prior to the commonly cited determinants of segregation of the 20th century.     

Gradients,Property Templates,and Land Use Change

The scarcity of historical individual-level data makes understanding historic land use decisions and the influence of physiochemical gradients on these decisions difficult. Here, we present several measures of early property agricultural quality based on commonly available data, including elevation and soil type. These analyses demonstrate the influence of physiochemical gradients on initial land division patterns in the Gywnns Falls watershed (Baltimore, Maryland). Moreover, we examine the influence of the template created by early property decisions on continuing human-driven landscape dynamics. This influence is illustrated by 1900-era forest cover patterns, park locations, and modern transportation networks.     

Spatial mining of migration patterns from web demographics

Volunteered Geographic Information, social media, and data from Information and Communication Technology are emerging sources of big data that contribute to the development and understanding of the spatiotemporal distribution of human population. However, the inherent anonymity of these crowd-sourced or crowd-harvested data sources lack the socioeconomic and demographic attributes to examine and explain human mobility and spatiotemporal patterns. In this paper, we investigate an Internet-based demographic data source, personal microdata databases publicly accessible on the World Wide Web (hereafter web demographics), as potential sources of aspatial and spatiotemporal information regarding the landscape of human dynamics. The objectives of this paper are twofold: (1) to develop an analytical framework to identify mobile population from web demographics as an individual-level residential history data, and (2) to explore their geographic and demographic patterns of migration. Using web demographics of Vietnamese–Americans in Texas collected in 2010 as a case study, this paper (1) addresses entity resolution and identifies mobile population through the application of a Cost-Sensitive Alternative Decision Tree (CS-ADT) algorithm, (2) investigates migration pathways and clusters to include both short- and long-distance patterns, and (3) analyze the demographic characteristics of mobile population and the functional relationship with travel distance. By linking the physical space at the individual level, this unique methodology attempts to enhance the understanding of human movement at multiple spatial scales.     

Residential Settlement Patterns among Immigrants in Atlanta Metropolitan Area,Georgia

This study investigates the diversity of immigrant populations and their spatial residential patterns at the census tract level within twenty‐nine counties of the Atlanta Metropolitan Statistical Area (MSA), using 2013 American Community Survey (ACS) five‐year estimates summary data from 2008–2012. Correlations coefficient analyses with local job opportunities, education level, and English proficiency and the presence of foreign‐born populations are also discussed. We find that immigrants have different patterns of settlement depending on their place of birth and region of origin. Local job opportunities, dominant language at home, and education levels also relate to the distribution of foreign‐born populations. Further research opportunities concerning differential settlement are discussed.     

The development and application of a land use diversity index for Oklahoma City,OK

Characteristics of land use and land use diversity play an integral part in determining how a city develops, and also can influence the allocation of funds for urban redevelopment, crime prevention, and civic and transportation projects. With the rise of the sustainability movement and concern about environmental degradation, more cities are interested in optimizing transportation routes and utilizing their budgets to accommodate high-priority projects that benefit the most people. This paper develops a method for measuring land use diversity for Oklahoma City, Oklahoma, and explores the relationship between land use diversity and a suite of demographic variables. To accomplish this, a land use diversity index (LUDI) was developed to examine spatial patterns in diversity of land use. The analysis shows that areas with a higher LUDI are near interstates and main corridors while areas with a lower LUDI are located closer to where newer residential areas have been developed. A multivariate statistical approach was then used to identify those variables that might explain the specific spatial patterns of the LUDI in Oklahoma City. The results identify the relationship between the demographic variables and land use across the city, and illustrate the most interesting and significant spatial patterns. By coupling the land use diversity index and demographic data, Oklahoma City planners can apply the information to help make important decisions for the city, and it will provide quantifiable data to support their decisions.     

Analyzing the distribution of human activity space from mobile phone usage: an individual and urban-oriented study

Travel activities are embodied as people’s needs to be physically present at certain locations. The development of Information and Communication Technologies (ICTs, such as mobile phones) has introduced new data sources for modeling human activities. Based on the scattered spatiotemporal points provided in mobile phone datasets, it is feasible to study the patterns (e.g., the scale, shape, and regularity) of human activities. In this paper, we propose methods for analyzing the distribution of human activity space from both individual and urban perspectives based on mobile phone data. The Weibull distribution is utilized to model three predefined measurements of activity space (radius, shape index, and entropy). The correlation between demographic factors (age and gender) and the usage of urban space is also tested to reveal underlying patterns. The results of this research will enhance the understanding of human activities in different urban systems and demographic groups, as well as providing novel methods to expand the important and widely applicable area of geographic knowledge discovery in the age of instant access.     

A Proposed Method for Spatial Data Disaggregation and Interpolation

The Model for External Reliance of Localities in (MERLIN) Coastal Management Zones is a proposed solution to allow scaling of variables to smaller, nested geographies. Using a principal components analysis and data normalization techniques, smaller scale data relationships are linked to data relationships within larger geographies and scaling factors are developed for each. In using these relationships, MERLIN will derive a set of variables scaled to a desired geography representing the attributes of the population present. The concept driving MERLIN's development is that correlations between variables at a specific location and level of geography can be tested and applied to other scales for interpolations. This flexibility and relative simplicity of calculation make MERLIN a valuable tool in many types of social research and for data interpolation. In a pilot study, the MERLIN model is applied to counties along the Gulf Coast of Florida to create a well-being index at the census tract level. Results reveal a landscape that better accounts for the localized patterns of values situated within a larger geographic unit.     

Guest editorial: the 2001 UK census: remarkable resource or bygone legacy of the 'pencil and paper era'?

National censuses are expensive. They are conducted infrequently. They collect information that some feel infringes their human rights, and people are required by law to complete them. The outputs are not perfect, and in some situations may be misleading. Some suggest that censuses hark back to a period when regularly collected administrative data were not available. These are some of the views held about national censuses. Why, then, would others argue that they are an essential resource? In this paper, we consider some of the pros and cons of conducting national censuses, before introducing a series of papers that draw on early data available from the 2001 UK census. We argue that these papers, and the wealth of research that will be conducted in the future with 2001 census data, make a strong case for supporting the compulsory collection of personal information about the 'entire' population every ten years.     

Changes in Correlation Coefficients with Spatial Scale and Implications for Water Resources and Vulnerability Data

Previous studies of correlation coefficients between paired observations using census, hydrologic, and remote sensing data abound. It is well established that bivariate relationships at coarser spatial resolutions are often stronger than at finer resolutions. No assessment as yet, however, corroborates this tendency with water resources variables. In this study, multiscale correlations between water use or water availability and population are presented in three river basins—the Missouri (United States), Danube (Europe), and Ganges (South Asia). High-resolution gridded data sets were obtained at 0.5° and resampled to fourteen different geographic scales to examine the effects of scale on the strength and trends of correlations. Correlation coefficients between most variable pairs increased at coarser scales. Smoothing fine-scale spatial patterns in the data at coarser scales is posited as a possible explanation. The increase was not often linear, however, nor was there always an increase. The Missouri Basin did not show a significant increase in correlations between water use and population with grid-cell size and nonlinear increases are evident in the Ganges Basin.     

Coverage inequality and quality of volunteered geographic features in Chinese cities: Analyzing the associated local characteristics using geographically weighted regression

The volunteered geographic information (VGI) gains increasing popularity with the general public and scientific community. However, the optimism about the VGI has been tempered by two critical issues: inequality in data coverage (social justice) and data quality. It therefore requires a better understanding of the mechanism driving VGI contributions and content quality. With a case of China, this paper demonstrates one potential avenue, examining the associations between VGI coverage/quality and local demographic and socioeconomic characteristics. In particular, VGI data are harvested from the OpenStreetMap for 333 cities in China. VGI coverage is measured by the total volume of different geographic features (point, line and polygon); and VGI quality is described from two aspects: completeness and accuracy. Geographically weighted regression (GWR) shows that both demographic and socioeconomic factors have statistically significant influences on VGI coverage and quality. More specifically, densely populous cities with more young, educated and non-agricultural people enjoy higher VGI coverage and quality. Cities with lower VGI coverage and quality are primarily located in the western and southwestern regions where the ethnic minorities concentrate. High VGI coverage and quality are possibly observed in economically developed cities with high marketization degree. Besides, possibility of high VGI coverage and quality occurs in cities with more labor in scientific research and greater percentage of employers in the tertiary industry. The GWR also demonstrates that the strength and nature of the obtained relationships vary across the 333 cities. The spatial non-stationary relationships may partially answer for the controversial empirical conclusions in earlier case studies at different scales. Quantitative analysis (Gini index, Lorenz curve and Moran's I index) further evidences the great inequality in VGI coverage and quality. It can be safely inferred that the differences in engagement and use of VGI, as a new digital divide, can raise troubling concerns on the social justice implications.     

On the Components of Indigenous Population Change

The 2001 census count of Indigenous Australians produced an intercensal change in numbers that cannot be explained by demographic processes alone. Using census and vital registration data, this paper unravels the components of such change and provides new insight into Indigenous population dynamics. In particular, it establishes the first estimates of proximate determinants of fertility, and extends mortality analysis by examining the components of low Indigenous life expectancy. Results show that demographic factors account for only 69 per cent of population change. Of these components, national Indigenous fertility is found to be below replacement level, while lack of convergence between Indigenous and non-Indigenous mortality remains. As Indigenous socio-economic circumstances are spatially diverse, the paper also explores the geography of demographic processes using data for 36 ATSIC regions and capital city/balance of State classifications. This reveals continued high fertility across parts of north Australia and an indication that mortality levels are associated with degree of urban residence.     

The Social Geographies of Indigenous Population and Housing in Australia's Regional Urban Centres

In recent years, social scientists have turned increasing scholarly focus to the ‘post-colonial city’, as it relates to the lived experiences and socio-cultural worlds of urban-dwelling Indigenous peoples. In Australia, this literature is concerned primarily with issues of identity, power and representation in big cities. More conventional geographic analyses regarding socio-spatial, economic and demographic aspects of urban Indigenous experience have been mostly absent from this discourse. In this paper, we begin to address this gap within the literature by identifying notable socio-spatial, economic, and demographic features of urban Indigenous experience in regional Australian cities. We draw on census and administrative data to empirically examine Indigenous residency, presence, and uneven housing market access in regional urban centres. The analysis builds a national comparative picture regarding population change, tenure type, crowding and housing affordability in these localities. At the same time, it ‘drills down’ to explore some of the drivers and implications of change against these indicators in one case study locality: Geraldton Western Australian. This case study analysis, based on recent fieldwork, grounds the broad identified trends in a localised narrative, illuminating some of general and specific socio-economic and socio-spatial dynamics that shape urban life for Indigenous peoples in regional Australia and the implications of these dynamics for broader questions of co-existence between Indigenous and non-Indigenous Australians.