Spatial Analytics — A Missing Key to Ending
Homelessness
Completed Research
Hafsa Aasi
Claremont Graduate University
[email protected]
Yoonmi Lee
Claremont Graduate University
[email protected]
Abstract
Homelessness is a complex social phenomenon which requires a nuanced understanding of the development and implementation of lasting solutions. Such an endeavor requires inter-disciplinary research and collaboration. Yet the body of scholarly work on this phenomenon is invariably compartmentalized by distinct disciplines, often being studied separately in urban studies, sociology, public policy, health, and economic literature. It is no surprise then that there is a dearth of scholarly work on homelessness focused on the importance of understanding its spatial dimension. The studies that do include spatial data on homelessness tend to limit their analysis efforts to basic map visualizations. This paper contributes to Geographic Information Science by surveying the peer-reviewed research literature to assess current understanding of the spatial dimension of homelessness, and how this understanding informs efforts to address homelessness, to uncover the gaps that remain.
Keywords
GIS, spatial analysis, homelessness, humanitarian technology.
Introduction
Homelessness is a complex social phenomenon which requires a nuanced understanding of the development and implementation of lasting solutions. Such an endeavor requires inter-disciplinary research and collaboration. Yet the body of scholarly work on this phenomenon is invariably compartmentalized by distinct disciplines, often being studied separately in urban studies, sociology, public policy, health, and economic literature. It is no surprise then that there is a dearth of scholarly work on homelessness focused on the importance of understanding its spatial dimension. The studies that do include spatial data on homelessness tend to limit their analysis efforts to basic map visualizations. This paper contributes to Geographic Information Science by surveying the peer-reviewed research literature to assess current understanding of the spatial dimension of homelessness, and how this understanding informs efforts to address homelessness, to uncover the gaps that remain.
The first section of this paper lays out the conceptual foundation for understanding the spatial dimension of the social phenomena in general. The second section of this paper describes the research methodology and data collection for this study, focusing on homelessness. The third section of this paper surveys the peer-reviewed research on homelessness focused on developing an understanding of the spatial dimension of the phenomenon. The fourth section of this paper analyzes and synthesizes the scholarly works reviewed. The fifth section of this paper highlights the gaps that remain in the research to better understand and address the spatial dimension of homelessness and, in doing so, adds to the body of scholarly work on spatial analysis and Geographic Information Science. In this study, Geographic Information Systems (GIS) refer to the technologies that support research and practice in Geographic Information Science.
CORE Metadata, citation and similar papers at core.ac.uk
Problem Statement
Homelessness researchers with limited or no GIS expertise need domain specific guidance on the functionalities they should consider using to understand and address the spatial dimension of homelessness in their studies. For example, efforts to end homelessness can be better understood by examining the spaces and places where it exists, and the accessibility of supportive resources such as shelters, food banks, and job training. Unfortunately, many homelessness researchers have limited, if any, GIS expertise to perform such spatial analyses. This is because “(s)patial analytic functionality is the nucleus of GIS software, but many existing analysis tools are hard to be identified and used for non-expert users without specific training of GIS software skills because of the complex hierarchical organization of functionalities (Gao & Goodchild 2013).”
Opportunity Statement
“Organizations that leverage their increasing volume of geospatial data have the potential to enhance their strategic and organizational decisions (Erskine et al. 2014).” Whether such an organization is in academia, the public sector, an NGO, or a business, the use of GIS and spatial analysis can add another dimension to the understanding of complex, seemingly intractable problems such as homelessness. This paper seeks to uncover and promote such opportunities in academia to understand and address the spatial dimension of homelessness. For example, insights from this paper can help inform homelessness research on the spatial distribution of the homeless, their networks, their mobility, their access to resources, targeted outreach to them by service providers, and site suitability of supportive resources.
Conceptual Foundation
In order to assess research efforts aimed at understanding the spatial dimension of homelessness, it is important to begin by taking inventory of existing approaches, theories, and techniques to do so in general for any social phenomenon. GIS advocates, both researchers and practitioners, have been trying to come up with ways to help novice and non-expert users of GIS software answer a variety of different types of spatial questions. Most of these attempts categorize different spatial analysis functionalities. Some researchers have even attempted to offer actionable recommendations for conducting spatial analysis. Several ways of categorizing spatial analysis functionalities were considered for the conceptual foundation of this paper.
Nearly four decades ago, Openshaw predicted that Geographic Information Systems technology will have increasingly varied applications but is in danger of being limited to use for mapping rather than for spatial analysis that can enrich our understanding of the role location plays in highly focused scientific inquiry (Openshaw 1990).
Albrecht attempted to address the concern of GIS use primarily being limited to mapping by taking a task-oriented approach to categorizing GIS functionality independent of the structure of the data and the applications they are used in. He evaluated six categories of spatial analysis methods, which include Search, Locational Analysis, Terrain Analysis, Distribution / Neighborhood, Spatial Analysis, and Measurements to determine how universal their usefulness is for performing analytical GIS functions. He argued that most applications only need at most 20 universal analytical GIS operations (Albrecht 1998). These categories are too broad for most homelessness researchers who are non-expert GIS users to be able to identify specific spatial analysis techniques from.
Some researchers proposed a new semantic GIS user interface that is based on spatial questions comprising a question layer, a semantic layer, an implementation layer, a spatial concept ontology, a domain ontology, and a task ontology along with sample spatial questions categorized by Search / Location and Extent, Data Basics / Processing / Conversion, Distributions / Patterns / Neighborhood, Relations / Associations, Terrain / Surface, and Time. Each category includes sample spatial questions and the GIS functionality that can be used to answer those questions (Gao & Goodchild, 2013). While this approach holds promise in the future, the semantic GIS user interface proposed is not yet available for homeless researchers with little to no GIS expertise.
More recently, a synthesis of an anthology of research on the practice of spatial analysis by another research team identified six categories of spatial analysis methods and techniques for complex socio-spatial problem-solving including socio-spatial interaction models (SIM), socio-spatial optimization models (SOM), spatial statistics (SSt), spatial econometrics(SE), geo-computation(GC), and spatial simulation(SSim) (Briassoulis, Kavroudakis, & Soulakellis 2019). However, these categories are still too broad for most homelessness researchers who are non-expert GIS users to be able to identify specific spatial analysis techniques from.
Esri, currently one of the largest commercial GIS technology vendors and evangelist for GIS technology at the time of this writing, has been making its own efforts to increase the use of GIS functionality, understandably by increasing the adoption of their GIS software. In addition to providing technical support, training resources, and massive open online courses, they have recently offered a clearer path to performing increasingly advanced spatial analysis. To help GIS technology users, they have developed a general spatial analysis framework organized by six main categories comprising a total of 26 high-level spatial analysis questions (Cappelli 2013). This framework is clearly aimed at helping non-expert and novice GIS users more easily identify the functionality needed to answer a variety of different spatial questions. Using this framework, this study reviews the literature on homelessness to uncover patterns of GIS technology used by researchers from different disciplines in order to determine the extent to which they have advanced our understanding of the spatial dimension of homelessness. This framework informs the research methodology of this study, as described in the next section of this paper.
Methodology
The research methodology of this study is predicated on the conceptual foundation laid out in the previous section of this paper. As such, the scholarly literature reviewed in this study focuses on those peer-reviewed journal articles that attempt to understand and/or address homelessness by incorporating spatial analysis in their research methodologies. Since the aim of this paper is to uncover the gaps that remain in understanding the spatial dimension of homelessness, the literature reviewed is analyzed to determine which of the research articles use spatial analysis techniques corresponding to the spatial analysis categories and questions from Cappelli’s framework.
Data Collection
While the definitions of homelessness vary based on different parameters and agendas, this paper is not focused on exploring their nuances. This particular literature review is focused on the scholarly research on homelessness which uses GIS and/or spatial analysis techniques. Multiple sources were used as part of the search strategy in order to conduct a systematic narrative overview of the literature.
Google Scholar, which in recent years has caught up to Web of Science and Scopus in terms of academic literature coverage and search capability (Harzing & Alakangas 2016), was used to search multiple bibliographic databases for peer-reviewed articles published in English. The search strings queried were “GIS, homelessness” and “Spatial Analysis, homelessness.” Initially, the first search string yielded about 6,970 results and the second search string yielded 37,000 results. Those search results, the primary focus of which was not homelessness, were excluded from this review along with papers the primary analysis methods of which did not include GIS nor spatial analysis techniques. Dissertations, theses, government reports, non-profit reports, slide decks, student assignments, books and white papers were also excluded since the criteria for their peer-review differ from those for peer-reviewed scholarly journal and conference papers. Finally, duplicate results for the same paper, otherwise known as stray citations, were also excluded from this literature review. Each scholarly study that was included was reviewed and the spatial analysis method(s) and finding(s) of the study were collected in an organized manner.
Analysis
The six categories of spatial analysis and the corresponding 26 questions from Cappelli’s framework have been used to code and analyze the data. Since they are general spatial analysis categories of spatial analysis techniques, they have been adapted to focus on understanding and addressing homelessness using Cappelli’s language to:
“understand where (homelessness) exists
measure the size, shape, and distribution of (homelessness)
determine how locations where (homelessness) is occurring are related find out the best locations and paths to address (homelessness)
detect and quantify patterns of (homelessness) make predictions about (homelessness)”
Category 1, understand where homelessness exists, is represented as C1.
Category 2, measure the size, shape, and distribution of homelessness, is represented as C2.
Category 3, determine how locations where homelessness is occurring are related, is represented as C3. Category 4, find out the best locations and paths to address homelessness, is represented as C4.
Category 5, detect and quantify patterns of homelessness, is represented as C5. Category 6, make predictions about homeless, is represented as C6.
The 26 spatial analysis questions are numbered and coded as Q1, Q2, Q3, and so on. The spatial analysis questions corresponding to C1 are represented as Q1, Q2, and Q3. The spatial analysis questions corresponding to C2 are represented as Q4 and Q5.
The spatial analysis questions corresponding to C3 are represented as Q6, Q7, Q8, Q9, and Q10. The spatial analysis questions corresponding to C4 are represented as Q11, Q12, Q13, Q14, and Q15. The spatial analysis questions corresponding to C5 are represented as Q16, Q17, Q18, and Q19.
The spatial analysis questions corresponding to C6 are represented as Q20, Q21, Q22, Q23, Q24, Q25, and Q26.
Q1 has been adapted using Cappelli’s language to understand “where (homelessness and supportive resources) are.”
Q2 has been adapted using Cappelli’s language to understand “where the variations and patterns in values (of homelessness and supportive resources) are.”
Q3 has been adapted using Cappelli’s language to understand “where and when locations and values (of homelessness and supportive resources) change.”
Q4 has been adapted using Cappelli’s language to calculate “individual feature geometries (of homelessness and supportive resources).”
Q5 has been adapted using Cappelli’s language to calculate “geometries and distributions of (homelessness and supportive resources) feature collections.”
Q6 has been adapted using Cappelli’s language to determine “what is nearby or coincident (to homelessness and supportive resources).”
Q7 has been adapted using Cappelli’s language to determine and summarize “what (homelessness and supportive resource) is within an area.”
Question 8 has been adapted using Cappelli’s language to determine “what is closest (to homelessness and supportive resources).”
Q9 has been adapted using Cappelli’s language to determine “what (homelessness and supportive resource) is visible from a given location.”
Q10 has been adapted using Cappelli’s language to determine “overlapping (homelessness and supportive resource) relationships in space and time.”
Q11 has been adapted using Cappelli’s language to find “the best locations that satisfy a set of criteria (for homeless supportive resources).”
Q12 has been adapted using Cappelli’s language to find “the best allocation of (supportive) resources (for the homeless) to geographic areas.”
Q13 has been adapted using Cappelli’s language to find “the best route, path, or flow along a network (for the homeless and supportive resources).”
Q14 has been adapted using Cappelli’s language to find “the best route, path, or corridor across open terrain (for the homeless and supportive resources).”
Question 15 has been adapted using Cappelli’s language to find “the best supply locations given known demand and a travel network (for the homeless and supportive resources).”
Q16 has been adapted using Cappelli’s language to detect where “the significant (homelessness and supportive resources) hot spots, anomalies, and outliers” are.”
Q17 has been adapted using Cappelli’s language to detect “the local, regional, and global (homelessness and supportive resources) spatial trends.”
Q18 has been adapted using Cappelli’s language to detect “which (homelessness and supportive resources) features/pixels are similar, and how can they be clustered, classified, and identified.”
Q19 has been adapted using Cappelli’s language to detect if “(homelessness and supportive resource) spatial patterns (are) changing over time.”
Q20 has been adapted using Cappelli’s language to “(identify, rank, and predict) similar locations (of homelessness and supportive resources) given a success case.”
Q21 has been adapted using Cappelli’s language to find “the factors that explain observed spatial patterns and make predictions (about homelessness and supportive resources).”
Q22 has been adapted using Cappelli’s language to “(interpolate) a continuous surface and trends from discrete sample observations (of homelessness and supportive resources).”
Q23 has been adapted using Cappelli’s language to predict “how and where (homelessness and supportive resource) objects spatially interact.”
Q24 has been adapted using Cappelli’s language to predict “how and where (homelessness and supportive resource) objects affect wave propagation.”
Q25 has been adapted using Cappelli’s language to predict “where (homelessness and supportive resource) phenomena will move, flow, or spread.”
Q26 has been adapted using Cappelli’s language to predict “(homelessness and supportive resource) what-if.”
Table 1. Spatial Analysis Methods Used to Study Homelessness
Author(s), Year Spatial Analysis Method(s) Used Category Question 1 Almquist, 2020 Spatial Bernoulli Graphs & Gravity Models C6 Q23 2 Byrne et al., 2014 Mapping
(super-imposing centroids)
C1 C2
Q1 Q4 3 Cloke et al., 2001 Thematic Mapping C1 Q2 4 Cook & Corbett,
2019 Group Participatory Mapping C1 Q1 5 Curtis et al., 2015 Spatial Video Geonarrative (SVG) Mapping C1 Q1 6 Fiedler et al.,
2006 Dasymetric Mapping C1 Q1
7 Gawron et al.,
2017 Mapping Drive Times
Cluster and Outlier Analysis
C1 C3 C5 Q1 Q8 Q16 8 Iwata & Karato,
2011
Spatial Regression C6 Q21
9 Kaplan et al., 2019 Geocoding and Mapping Mean center
Standard Distance Point Density
Resource Proximity Using Spatial Joins
C1 C2 C3 C1 C3 Q1 Q5 Q6 Q2 Q7 10 Leung et al. 2015 Geocoding and Mapping
Buffers
C1 C2
Q1 Q5
2005 Proximity – Spatial Query Thematic Mapping C3 C1 Q6 Q2 12 Marr et al., 2009 Geocoding and Mapping
Cluster Analysis C1 C5 Q1 Q16 13 Meert & Bouorgeois, 2005 Mapping C1 Q1
14 North et al., 2017 GPS Mapping C1 Q1
15 Rukmana, 2010 Geocoding
Mean Center & Standard Deviational Ellipse Hotspot Analysis – Single Kernel Density
C1 C2 C5 Q1 Q5 Q16 16 Simon et al. 2019 Mapping
Measuring Activity Size
C1 C2
Q1 Q5 17 Suzuki, 2008 Thematic Mapping
Spatial Regression C1 C6 Q2 Q21 18 Townley et al.
2016 Geocoding and Participatory Mapping Activity Spaces and SDE C1 C2 Q1 Q5 19 Williams &
Sheehan, 2015 Socio-Spatial Daily Paths N/A N/A 20 Wong & Hillier,
2001 Geocoding Thematic Mapping Location Quotients C1 C1 N/A Q1 Q2 N/A Table 1. Spatial Analysis Methods Used to Study Homelessness cont.
Results
The findings from this study are expected to reveal the gaps in research employing spatial analysis techniques to examine issues surrounding homelessness including but not limited to access to community resources, healthcare, and food in which spatial context is important in understanding and problem-solving.
Category 1 had 20 matches, mostly for mapping.
o Question 1 had 15 matches for spatial analysis techniques including: geocoding, location, GPS, participatory, and dasymetric mapping. o Question 2 had 5 matches for spatial analysis techniques including:
thematic and dot density mapping. o Question 3 had 0 (no) matches at all.
Category 2 had 6 matches, some for buffers and some for spatial descriptive statistics. o Question 4 had 1 match for the spatial analysis technique:
super-imposing centroids.
o Question 5 had 5 matches for spatial analysis techniques including:
mean center, standard deviational ellipse, activity size, and buffers. Category 3 had 4 matches for drive times, standard distance, spatial joins, and spatial queries.
o Question 6 had 2 matches for spatial analysis techniques including: standard distance and spatial query.
spatial joins.
o Question 8 had 1 match for the spatial analysis technique: drive times.
o Question 9 had 0 (no) matches at all. o Question 10 had 0 (no) matches at all. Category 4 had no matches at all.
o Question 11 had 0 (no) matches at all. o Question 12 had 0 (no) matches at all. o Question 13 had 0 (no) matches at all. o Question 14 had 0 (no) matches at all. o Question 15 had 0 (no) matches at all.
Category 5 had 3 matches for hot spot, cluster, and outlier analysis.
o Question 16 had 3 matches for spatial analysis techniques including: single kernel density, hotspot, cluster and outlier analysis. o Question 17 had 0 (no) matches at all.
o Question 18 had 0 (no) matches at all. o Question 19 had 0 (no) matches at all.
Category 6 had 3 matches for spatial regression and spatial interaction gravity models. o Question 20 had 0 (no) matches at all.
o Question 21 had 2 matches for spatial analysis techniques including: spatial regression.
o Question 22 had 0 (no) matches at all.
o Question 23 had 1 match for the spatial analysis technique: gravity models.
o Question 24 had 0 (no) matches at all. o Question 25 had 0 (no) matches at all. o Question 26 had 0 (no) matches at all.
Present Understanding of the Spatial Dimension of Homelessness
Much of the GIS use for spatial analysis in the literature reviewed, 85%, falls under Category 1 of Cappelli’s framework, to understand where homelessness exists. The spatial analysis techniques corresponding to this category require very little GIS expertise. In fact, one of the studies reviewed even touts ease of implementation as a benefit of the spatial analysis techniques researchers integrate into their spatial video geo-narratives to complete “case studies on post-disaster psychopathology, crime, mosquito control, and TB in homeless populations” (Curtis et al. 2015).
Only 30% of the literature reviewed uses GIS for spatial analysis techniques which fall under Category 2 of Cappelli’s framework, to measure the size, shape, and distribution of homelessness. The spatial analysis techniques corresponding to this category require more advanced GIS expertise than those corresponding to Category 1. This is exemplified by one of the studies reviewed that uses a mixed method approach to determine the factors that affect the activity space sizes of homeless men and women in two Czech cities, one large, Prague, and one small, Pilsen. They also conduct mobility interviews with them and mapped their GPS tracking data captured over a week-long period. They find that the activity spaces of the homeless seem to be most strongly related to the size of the city and then to a lesser extent, to transport infrastructure and services provision (Simon et al. 2019).
Another 15% of the literature reviewed uses GIS for spatial analysis techniques which fall under Category 3 of Cappelli’s framework, to determine how locations where homelessness is occurring are related. The spatial analysis techniques corresponding to this category also require more advanced GIS expertise than those corresponding to Category 1. This is exemplified by one of the studies reviewed that uses personal interviews and GIS to run spatial queries to help shelter providers and decision-makers. The study finds that shelter providers and decision makers often site new homeless shelters and services based on their perceptions of that system which are only partially empirically supported (Lobao 2005).
None of the literature reviewed falls under Category 4 of Cappelli’s framework, to find out the best locations and paths to address homelessness. The spatial analysis techniques corresponding to this
Another 15% of the literature reviewed uses GIS for spatial analysis techniques which fall under Category 5 of Cappelli’s framework, to detect and quantify patterns of homelessness. The spatial analysis techniques corresponding to this category also require more advanced GIS expertise than those corresponding to Category 1. This is exemplified by one of the studies reviewed that uses a mixed method approach to better understand the place-homeless survival nexus in Los Angeles County. This study uses in-depth life-history interviews of 25 homeless research participants from three types of urban spaces, selected using maximum variation sampling and cluster analysis. The study finds significant differences in the place-survival nexus across marginal, transitional, and prime urban spaces (Marr et al. 2009).
Another 15% of the literature reviewed uses GIS for spatial analysis techniques which fall under Category 6 of Cappelli’s framework, to make predictions about homelessness. The spatial analysis techniques corresponding to this category also require more advanced GIS expertise than those corresponding to Category 1. This is exemplified by one of the studies reviewed that uses the Spatial Auto-Regression Model to analyze a spatial distribution of the homeless people in Japan, finding that availability of employment and free or low-cost health care and food determine their spatial distribution there (Suzuki 2008).
Two of the spatial analysis techniques, socio-spatial daily paths and location quotients, used in the studies reviewed could not be neatly categorized by Cappelli’s framework so they were assigned the code N/A for not applicable. However, they were still included in the analysis since both of these spatial analysis techniques contribute to understanding the spatial dimension of homelessness. The use of socio-spatial daily paths was helpful in applying geographic theory to examine the lived experiences of geography, socio-spatial networks, and socio-spatial daily paths of homeless youth (Williams & Sheehan, 2015). The use of location quotients helped differentiate homelessness prevention service use in Philadelphia, Pennsylvania (Wong & Hillier 2001)
Conclusion
The findings of this literature review suggest that despite advances in spatial analysis approaches and techniques, decades later, Openshaw’s prediction still holds true, and GIS use in homelessness research is primarily limited to mapping. Understanding and ending homelessness calls for the much-ignored spatial dimension of homeless to be embraced and incorporated in current and new efforts facilitated by GIS to use advanced spatial analysis techniques well beyond mapping. For example, it requires dealing with the modifiable areal unit problem (MAUP), the activity space, accessibility, and even small area estimation. Each of these constructs needs to be operationalized at a level that is best suited to study the homeless population and supportive resources. It also requires understanding that the boundaries of the activity space for the homeless population may vary from person to person depending upon location, proximity to resources, social ties, cause of homelessness, age, health, and gender. Homelessness researchers with limited or no GIS expertise clearly need domain specific guidance on the functionalities they should consider using to understand and address the spatial dimension of homelessness in their studies. These functionalities include spatial analysis techniques corresponding to Category 2, 3, 4, 5, and 6 of Cappelli’s framework for spatial analysis.
Limitations of this Study
The data collection criteria for this study undoubtedly limit the sample size since much of the search results comprised of dissertations, theses, government reports, non-profit reports, slide decks, student assignments, books and white papers, none of which are analyzed in this review. The scope of this particular research endeavor does not include studies specifically investigating the effectiveness of policies in the public and private sectors. Such a focus in this literature review would detract from the intent of identifying the gaps in research per the scope of this paper, taking a detour into potentially normative conclusions which may be leveraged to influence political agendas of different organizations.
It is possible that the general search strategy for data collection in this paper might have missed literature that would only result in matches for specific spatial analysis techniques in the terms queried. Future research building on this study can include literature review resulting from a search of specific spatial analysis techniques best suited to understand the spatial dimension of homelessness.
GIS technology and spatial analysis techniques have not yet matured enough to meaningfully analyze heterogeneous populations in small areas. This limitation is noticeable in studies of complex social problems involving marginalized members of society. It is especially salient in studies of homelessness and the organizations involved in efforts to reduce its incidences. It is possible that some of the findings of this study are also influenced by these limitations.
It is also worth noting that this study does not explore whether or not the spatial analysis techniques used in the literature reviewed are well-suited to the studies that use them, due to time and funding constraints, potentially influencing some of the findings of this study. For example, in this paper, the standard deviational ellipse was used in several studies reviewed. While the standard deviational ellipse is useful for spatial description of a distribution, Yuill cautions that “(for) data sets with great eccentricity, the ellipse measure must be applied with great caution” (Yuill, 1971). However, Gong offers two theorems with proofs asserting that the standard deviational ellipse is not an ellipse but a closed curve which he names the standard deviational curve (Gong, 2002).
Implications for Future Research
The findings of this review have significant implications for homelessness researchers interested in understanding and addressing the spatial dimension of homelessness to improve and strengthen spatial literacy so that they can perform more advanced spatial analysis. The findings also highlight the need for a more comprehensive taxonomy of spatial analysis methods according to a framework like Cappelli’s that can serve as an actionable guide for homelessness researchers who are novice and non-expert GIS users. The findings also suggest the need for those researchers to list the spatial analysis techniques used in their studies as keywords to improve the ability for search engines to discover them. The findings can also inform new research endeavors in largely under-utilized categories of spatial analysis. For instance, a follow-up to this study can be done entailing the creation of a spatial accessibility index for resources to support the homeless for a given unit and location of analysis.
Acknowledgements
This research is made possible with the support of the first author’s doctoral dissertation committee, co-author, mentors, and colleagues.
REFERENCES
Albrecht, J. 1998. “Universal analytical GIS operations: A task-oriented systematization of data structure-independent GIS functionality,” Geographic Information Research: Transatlantic Perspectives, pp. 577-591.
Almquist, Z. W. 2020. “Large-scale spatial network models: An application to modeling information diffusion through the homeless population of San Francisco,” Environment and Planning B: Urban Analytics and City Science, (47:3), pp. 523-540. (https://doi.org/10.1177/2399808318785375) Briassoulis, H., Kavroudakis, D., & Soulakellis, N. 2019. “Introduction: The Practice of Spatial Analysis,”
in The Practice of Spatial Analysis: Essays in Memory of Professor Pavlos Kanaroglou, H. Briassoulis, D. Kavroudakis, & N. Soulakellis (eds.), pp. 1–18. (https://doi.org/10.1007/978-3-319-89806-3_1)
Byrne, T., Fargo, J. D., Montgomery, A. E., Munley, E., & Culhane, D. P. 2014. “The relationship between community investment in permanent supportive housing and chronic homelessness,” Social Service Review, (88:2), pp. 234-263. (https://doi.org/10.1086/676142)
Cappelli, C. 2013. The Language of Spatial Analysis, Redlands, CA: Esri Press.
Cloke, P., Milbourne, P., & Widdowfield, R. 2001. “The geographies of homelessness in rural England.” Regional Studies, (35:1), pp. 23-37. (https://doi.org/10.1080/00343400120025655)
Cook, S., & Corbett, J. 2019. ‘“Spatially, we are Creatures of Habit”–Understanding Gendered Geographies of Homelessness Using Participatory Mapping,’ Spatial Knowledge and Information Canada, (7:7), 4.
Curtis, A., Curtis, J. W., Shook, E., Smith, S., Jefferis, E., Porter, L., ... & Kerndt, P. R. 2015. “Spatial video geonarratives and health: Case studies in post-disaster recovery, crime, mosquito control and
tuberculosis in the homeless,” International Journal of Health Geographics, (14:1), 22. (https://doi.org/10.1186/s12942-015-0014-8)
Erskine, M. A., Gregg, D. G., Karimi, J., & Scott, J. E. 2014. “Business decision-making using geospatial data: A research framework and literature review,” Axioms, (3:1), pp. 10-30. (https://doi.org/10.3390/axioms3010010)
Fiedler, R., Schuurman, N., & Hyndman, J. 2005. “Improving census-based socioeconomic GIS for public policy: recent immigrants, spatially concentrated poverty and housing need in Vancouver,” ACME: An International E-Journal for Critical Geographies, (4:1), pp. 145-71.
Gao, S., & Goodchild, M. F. 2013. “Asking spatial questions to identify GIS functionality,” in 2013 Fourth International Conference on Computing for Geospatial Research and Application, pp. 106-110. IEEE. (https://doi.org/10.1109/COMGEO.2013.18)
Gawron, L., Pettey, W. B., Redd, A., Suo, Y., Turok, D. K., & Gundlapalli, A. V. 2017. ‘The “safety net” of community care: leveraging GIS to identify geographic access barriers to Texas family planning clinics for homeless women Veterans,’ in AMIA Annual Symposium Proceedings (2017), p. 750-759. American Medical Informatics Association.
Gong, J. 2002. “Clarifying the Standard Deviational Ellipse. Geographical Analysis,” (34:2), pp. 155–167. (https://doi.org/10.1111/j.1538-4632.2002.tb01082.x)
Harzing, A.-W., & Alakangas, S. 2016. “Google Scholar, Scopus and the Web of Science: a longitudinal and cross-disciplinary comparison,” Scientometrics, (106:2), pp. 787–804. (https://doi.org/10.1007/s11192-015-1798-9)
Iwata, S., & Karato, K. 2011. “Homeless networks and geographic concentration: evidence from Osaka city,” Papers in Regional Science, (90:1), pp. 27-46. (https://doi.org/10.1111/j.1435-5957.2010.00306.x)
Kaplan, A., Diver, K., Sandin, K., & Mill, S. K. 2019. “Homeless Interactions with the Built Environment: A Spatial Pattern Language of Abandoned Housing,” Urban Science, (3:2), p.65.
Leung, Z. C., Pun-Cheng, L. S., & Ho, A. P. 2015. “Locality service review and planning with GIS: A pilot study of spatial analysis of poverty data in Hong Kong,” Journal of Technology in Human Services, (33:1), pp. 38-52. (https://doi.org/10.1080/15228835.2014.998571)
Lobao, E. G., & Murray, A. T. 2005. “Exploratory analysis of the homeless shelter system in Columbus, Ohio. Geografiska Annaler: Series B,” Human Geography, (87:1), pp. 61-73.
Marr, M. D., DeVerteuil, G., & Snow, D. 2009. “Towards a contextual approach to the place–homeless survival nexus: An exploratory case study of Los Angeles County,” Cities, (26:6), pp. 307-317.
Meert, H., & Bourgeois, M. 2005. “Between rural and urban slums: A geography of pathways through homelessness,” Housing Studies, (20:1), pp. 107-125.
North, C. S., Wohlford, S. E., Dean, D. J., Black, M., Balfour, M. E., Petrovich, J. C., ... & Pollio, D. E. 2017. “A pilot study using mixed GPS/narrative interview methods to understand geospatial behavior in homeless populations,” Community Mental Health Journal, (53:6), pp. 661-671.
Openshaw, S. 1990. “Spatial analysis and geographical information systems: a review of progress and possibilities,” in Geographical Information Systems for Urban and Regional Planning, H. J. Scholten & J. C. H. Stillwell (eds.), pp. 153–163. (https://doi.org/10.1007/978-94-017-1677-2_14) Rukmana, D. 2010. “Gender differences in the residential origins of the homeless: Identification of areas
with high risk of homelessness,” Planning Practice & Research, (25:1), pp. 95-116.
Šimon, M., Vašát, P., Poláková, M., Gibas, P., & Daňková, H. 2019. “Activity spaces of homeless men and women measured by GPS tracking data: A comparative analysis of Prague and Pilsen,” Cities, 86, pp. 145-153. (https://doi.org/10.1016/j.cities.2018.09.011)
Suzuki, W. 2008. “What determines the spatial distribution of homeless people in Japan?” Applied Economics Letters, (15:13), pp. 1023–1026. (https://doi.org/10.1080/13504850600972394)
Townley, G., Pearson, L., Lehrwyn, J. M., Prophet, N. T., & Trauernicht, M. 2016. “Utilizing participatory mapping and GIS to examine the activity spaces of homeless youth,” American Journal of Community Psychology, (57:3-4), pp. 404-414. (https://doi.org/10.1002/ajcp.12060)
Williams, A., & Sheehan, R. A. 2015. “Emerging adulthood in time and space: The geographic context of homelessness,” Journal of Family Theory & Review, (7:2), pp. 126-143.
Wong, Y. L. I., & Hillier, A. E. 2001. “Evaluating a community-based homelessness prevention program: A geographic information system approach,” Administration in Social Work, (25:4), pp. 21-45.
Yuill, R. S. 1971. “The Standard Deviational Ellipse; An Updated Tool for Spatial Description,” Geografiska Annaler: Series B, Human Geography, (53:1), pp. 28–39. (https://doi.org/10.1080/04353684.1971.11879353)
