Geographic Crime Analysis

lable at ScienceDirect

Applied Geography 69 (2016) 65e74

Contents lists avai

Applied Geography

journal homepage: www.elsevier.com/locate/apgeog

Street profile analysis: A new method for mapping crime on major roadways

Valerie Spicer*, Justin Song, Patricia Brantingham, Andrew Park, Martin A. Andresen Institute of Canadian Research Studies, Simon Fraser University, Burnaby, BC, Canada

a r t i c l e i n f o

Article history: Received 10 November 2015 Received in revised form 16 February 2016 Accepted 21 February 2016 Available online 4 March 2016

Keywords: Crime mapping Environmental criminology Human movement Street profile analysis

* Corresponding author. E-mail addresses: vspicer@sfu.ca (V. Spicer), jdson

sfu.ca (P. Brantingham), apark@tru.ca (A. Park), andre

http://dx.doi.org/10.1016/j.apgeog.2016.02.008 0143-6228/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

Street profile analysis is a new method for analyzing temporal and spatial crime patterns along major roadways in metropolitan areas. This crime mapping technique allows for the identification of crime patterns along these street segments. These are linear spaces where aggregate crime patterns merge with crime attractors/generators and human movement to demonstrate how directionality is embedded in city infrastructures. Visually presenting the interplay between these criminological concepts and land use can improve police crime management strategies. This research presents how this crime mapping technique can be applied to a major roadway in Burnaby, Canada. This technique is contrasted with other crime mapping methods to demonstrate the utility of this approach when analyzing the rate and velocity of crime patterns overtime and in space.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

Modern cities are transforming at a fast pace and adapting to the changing demands of urban living. Developing multi-use buildings and meeting transportation needs while maintaining livability and public safety is a primary planning strategy for many urban centers (Loukaitous-Sideris, 2014; Newton, 2004; Skogan, 2015; Smith, Phillips and King, 2010). These competing infrastructures can sometimes create very specific crime dynamics that if left unat- tended over time alter, or in some cases contradict, the original planning concept for an area (Knapp, 2013; Spicer, 2012). The new crime analysis technique presented in this paper can be used to identify areas where crime surges along major roadways and to compare these patterns to transecting roadways. This mapping technique can clearly visualize temporal variances, crime type comparisons and historical crime trends.

Street profile analysis is ideal for small and linear places where conventional analytical approaches are not fully suitable for visu- alizing of crime in these spaces. Most often, practitioners use maps to visualize crime patterns such as kernel density maps and aggregate address count maps (Chainey & Ratcliffe, 2005; Chainey, Tompson, and Uhlig, 2008; Eck and Weisburd, 2005). These

g@sfu.ca (J. Song), pbrantin@ sen@sfu.ca (M.A. Andresen).

techniques are useful in presenting crime patterns throughout an area in order to expose crime hot spots and high crime locations. However, in order to demonstrate crime velocity or variance along a linear space, it may be preferable to engage in a graph approach, called street profile analysis, where the roadway is the x axis and crime count the y axis.

To the knowledge of the authors, this is a new crime mapping technique that can be utilized to study small urban areas along major roadways and to better understand the dynamics in these places. The research presented in this paper examines a major roadway in Burnaby, British Columbia. Burnaby in a jurisdiction in Metro Vancouver and the area under study contains several ele- ments including a large regional shopping centre, a mass trans- portation station, a major roadway, a bike path, businesses and multi-dwelling residences. Several street profile views of this place are presented to demonstrate the variety of crime dynamics and the utility of this new mapping technique. A transect meth- odology is used in conjunction to compare and contrast roadways that bisect this major roadway.

From a practitioner perspective, street profile analysis is “user friendly” and can be produced using most analytical packages. The advantage of this approach is that it can clearly define where crime specifically peeks, both in space and in time, thus optimizing pre- ventative strategies. Compared to techniques such as kernel density that diffuses the visual image of crime, this street profile technique sharpens the situation and can clearly demonstrate the problem. The street profile analysis is compared and contrasted to three

V. Spicer et al. / Applied Geography 69 (2016) 65e7466

other techniques. The strength and weaknesses of each technique is discussed.

2. Mapping framework

Environmental Criminology provides a theoretical framework for mapping crime in urban areas. Urban infrastructure and its impact on human movement and directionality influences crime occurrences by concentrating them into small, definable places. Crime analysis and mapping techniques can imbed these theoret- ical concepts into specific approaches that help to further define and understand these crime dynamics. The street profile mapping technique is based on these concepts of the urban infrastructure and is designed to demonstrate how crime occurs in small defin- able places and can surge due to specific dynamics in the environment.

2.1. City infrastructure

The urban infrastructure contains nodes, paths and edges where crime is concentrated (Brantingham & Brantingham, 1984). These are geographic spaces that also transition through temporal vari- ances creating definable crime patterns (Brantingham & Brantingham, 1984, 1993a, b). Nodes are places where human ac- tivity is concentrated such as the crossing of two paths or an attractive place such as a mall. The crime patterns at nodes should be viewed as temporal because the activity at these places is not generally consistent. As a simple example, malls are not usually open 24 h per day therefore and the potential for shoplifting is completely eliminated by the closure of the mall while this same closure creates the potential for burglary.

Paths are channels designed for human movement (vehicle e pedestrian e mass transportation e bicycle or foot paths). Edges are boundaries between places that transition from one type of place to another such as a single-family dwelling area to a commercial zone. Like nodes, paths and edges transition through various temporal states that impact crime patterns. Within this framework, the street network is of interest because it links and defines the interaction between these elements (Brantingham & Brantingham, 2015; Davies & Johnson, 2015; Johnson & Summers, 2015; Vandeviver, Van Daele, & Vander Beken, 2015).

In certain places in the urban environment these three elements are consolidated and in some ways compressed along certain street segments. This can create crime surges and the street profile analysis can locate these places, then assist in analyzing the tem- poral and crime dynamics. In particular, major roadways that contain activity nodes, high volume pathways and edges are sus- ceptible to these crime dynamics. Within this context, the street profile analysis can display the variance in crime density in a manner that clearly defines the impact of these three elements on crime patterns.

2.2. Effectively mapping small places

Crime place theory focuses on crime events in small places such as specific addresses, business types and block faces (Eck and Weisburd, 1995). These small places can be categorized by feature, cluster or facility (Eck and Weisburd, 1995). Features include aspects such as physical or social structure, while clusters can be understood as hot or cool spots, and facilities, or addresses, are places such as bars, problem premises, or parks (Eck and Weisburd, 1995).

Major roadways contain successive small places that create variability and sudden increases in criminal events along their trajectory. In a spatial analysis of street segments in Seattle, WA,

Groff, Weisburd and Yang (2010) found that contiguous street segments could have very different (sometimes opposite) trajec- tories. These increases or decreases in crime can be better under- stood using the elements defined in crime place theory (features e clustering e facilities). For instance, the presence of a facility like a mall on a major roadway produces criminogenic features such as reduced guardianship and increased target opportunity, and also creates a clustering of criminal events that may lead to small places next to one another having very different crime patterns. Another example is a strip of licensed establishments also generating a crime surge.

The street profile analysis can describe the linearity of a major roadway while at the same time exposing the multiple variances that can occur in such a place. In particular, this graph technique simplifies crime patterns and can produce comparisons on a single graph which allows for detailed analysis of crime, place and time.

2.3. Vizualizing the effect of crime attractors and crime generators

Crime attractors and crime generators are both small places with specific characteristics that make them higher crime areas (Brantingham & Brantingham, 1995). Crime generators are places that attract a large number of people such as a shopping or enter- tainment district, or a sporting venue. They produce crime because there are many people in attendance and also many potential tar- gets, thus the opportunity for crime is present, en masse. Crime attractors are also small places, however these are well-known for their criminal opportunities and, therefore, attract criminals. Strongly motivated offenders, usually not from that area, attend these places for criminal purposes. Some examples of crime attractors are drug or prostitution markets, or shopping malls near a major transit hub.

Crime patterns along major roadways may vary because of the number and size of crime attractors and generators they contain. Major roadways are linear spaces in the urban infrastructure that often bisect multiple neighborhoods. Crime peaks along these roadways, and their variance through time and crime type, can be better explained using the concepts of attractors and generators. As well, when considered longitudinally, the variation in crime peaks or the emergence of a crime surge may be the result of a generator turning into an attractor. The street profile analysis technique ex- poses crime attractors and generators by clearly defining crime density along the roadway.

2.4. Conceptualizing urban directionality

The relationship between urban directionality and crime has a long history founded on the concept of spatial criminology (Frank, Andresen, Cheng, & Brantingham, 2011; Rengert & Wasilchick, 1985). Research has demonstrated the influence of crime on macro urban directionality through the criminal attractiveness of town centers, the impact of mass transportation and the formation of criminogenic streets and neighborhoods (Herrman, 2013; Song, Spicer, Brantingham and Frank, 2013). The micro and individual aspect of directionality is explained by the geometry of crime (Brantingham & Brantingham, 1981). This perspective helps ex- plains and further clarify factors such as temporal constraint (Ratcliffe, 2006), directional bias by crime type (Van Daele & Bernasco, 2012), and more recently the directional bias of repeat property offender within a large-scale sample (Frank, Andresen, & Brantingham, 2012; Frank et al., 2011).

The analysis of major roadways is a meso analysis of urban directionality. Within large metropolitan cities there are smaller sub-sets of areas and pathways where human activity is concen- trated for various reasons. These may include attractive pedestrian

V. Spicer et al. / Applied Geography 69 (2016) 65e74 67

areas, shopping strips, an area known for pubs and restaurants, business districts, or a college campus. The street profile mapping technique allows researchers and practitioners to further under- stand the impact of these factors on crime patterns along major roadways. This technique also lends itself to comparative analysis between crime density and other factors such as vehicle or pedestrian traffic.

3. Research study

3.1. Study area

Fig. 1 is the study area and major roadway called Kingsway runs through this area from west to east. This arterial street traverses diagonally three major municipalities in the Metro Vancouver re- gion (Vancouver e Burnaby e New Westminster). In some portions of this roadway, a Skytrain route runs parallel to Kingsway. The Skytrain is a light-rail mass transit metro route that is mostly elevated above ground and services the Metro Vancouver region. The study area also includes a bike path that runs parallel to Kingsway. At the center of the study area is a regional shopping centre. This shopping centre is the largest mall in British Columbia. There are business towers attached as well as high-density dwell- ing residences surrounding this mall. The transecting roadways in this study area are mostly collector streets except for Royal Oak that is a minor arterial street servicing Burnaby. Two transecting

Fig. 1. Stud

roadways e Willingdon Ave and Royal Oak Ave e are highlighted in Fig. 1

3.2. Data

This study utilizes data from the Police Information Retrieval System (PIRS) and GIS Innovation data.

3.2.1. PIRS The Crime Data-Warehouse (CDW) is a collection of datasets

that contains officially reported crime events for Royal Canadian Mounted Police (RCMP) jurisdictions in British Columbia. RCMP jurisdictions vary in size of police membership and also area covered. This dataset contains approximately 4.4 million crime events. The study area is located within the jurisdiction of Burnaby RCMP. There are 38,855 crime events from the middle of 2001 to the middle of 2006 in the study area. The crime events are reported offences to the Burnaby RCMP. These events are varied including, but not limited to, property crime, violent crime, drug and traffic offences. These data contain attributes about the crime event such as date, time, location, offender information, and specific crime type.

3.2.2. GIS innovations data The 2006 road network data from a company named GIS In-

novations were used to geocode crime event locations. The data

y area.

V. Spicer et al. / Applied Geography 69 (2016) 65e7468

were interpolated to a 98.8% geocoding success rate. This road network data were also used to visualize the output results.

3.3. Mapping methodology

Five mapping techniques are compared to demonstrate the utility of the new technique proposed in this study. The first three are often used for crime analysis: kernel density, aggregate count to address and aggregate count to street segment (Chainey & Ratcliffe, 2005; Weisburd, Groff, & Yang, 2012). These techniques visualize crime using a map. The proposed street profile methodology pre- sents spatial data in an abstract format on a graph. This technique is beneficial when studying major roadways because it lends itself well to temporal and crime comparison analysis. As well, when merged with the transect mapping methodology, crime distribu- tion on adjacent and transecting roadways further amplifies the crime patterns on the major roadway.

3.3.1. Kernel density The kernel density function is used in a first instance to visualize

the data in this study. The search radius was set for three different distances: 50, 100, and 250 m. In all three instances, the maps were produced using 50 m rasters. A 50-m raster size was selected because this distance covers on average a half block. Therefore, this raster size shows variation at the block level.

3.3.2. Aggregate count to address This technique aggregates crime to specific addresses. Then

further classes of aggregation are formed to show high and low crime locations. Those crime locations that contain one to three crime incidents were treated with a slight random perturbation to ensure de-identification for privacy purposes and does not affect the visualization of the results.

3.3.3. Aggregate count to street segment This technique is a more recent development in crime analysis.

Both Weisburd et al. (2012) and Curman, Andresen, and Brantingham (2015) demonstrate the utility of this analysis spe- cifically when looking at historical crime patterns. In this tech- nique, crime count is aggregated to the street segment and then further classes of aggregation can be formed to show high crime street segments.

3.3.4. Street profile Unlike the three previous methods, the street profile method is

presented on a graph and used to study areas in a different manner to provide another description of the crime problem. The street profile is created using successive circular buffers that have a 50- m radius, overlapped at the center point, and aligned with the roadway. Fig. 2 illustrates the location of the buffers along the roadways and how these are overlapped in order to consolidate the crime that is shown in the street profile.

Once these data are collated, the output is converted to a line graph and can be exported to Excel and made into a chart. Trans- ecting streets can be labeled on the vertical axis to help orient the viewer.

3.3.5. Line-transect methodology Line-transect methodology is most often used in ecological

sampling for animals or plants (Manly & Navarro Alberto, 2015). Lines are placed through the study area in order to establish sys- tematic sampling methodology. In this study, we adapt this approach to the street network in order to analyze patterns of crime on the streets that transect the major roadway. When working with the street profile method, the line-transect methodology reveals

the condensed and directional nature of crime patterns and how transecting streets have alternative dynamics. We further add cir- cular 50 m buffers to demonstrate crime directionality through a static visualization. The direction of the buffers is angled in order to encompass both sides of each street.

4. Results

The crime events in this study are analyzed and visualized using the four methods: kernel density, aggregate to address, aggregate to street segment and street profile. These visualizations are dis- cussed in terms of their utility and limitations.

4.1. Kernel density

This first method utilizes the kernel density function. This is a common technique used in crime analysis and typically produces hotspot maps. In these examples, the study area is quite small therefore the pixelization is very pronounced. More often, the hotspot maps produced with this technique are of larger areas and the pixelization is more smoothed. Such representations can be problematic. When producing a value for each kernel, the kernel density method uses a bandwidth to capture the number of events within a specified area and then applies a spatial average (Bailey & Gatrell, 1995). Though it may be true that most users of kernel density functions are aware of this limitation, not all of those who interpret the resulting maps will be. Three different search radii were utilized to create the maps in Fig. 3 and are displayed using 50 m rasters.

The map that utilizes 50-m search radius for single crime events in Fig. 3 produces a confusing result in that there appears to be great variation within the study area. This variation may also lead to false conclusions about the actual location of crime hotspots (Song, Frank, Brantingham, & LeBeau, 2012). The inherent smoothing ef- fect of the kernel density function can actually create a hotspot between two crime locations rather than showing the reality of the situation because of the bandwidth and spatial averaging of the function as mentioned above (Song et al., 2012). As the search radius is increases to 100 m and 250 m in Fig. 3, the hotspot be- comes more generalized. Overall, the kernel density function is best used to provide a broad idea of crime and to locate high crime areas. However, in order to understand the specific location and dynamics of crimes, other techniques are necessary.

4.2. Aggregate count to address

This second method is also commonly used in crime analysis. In Fig. 4 crimes are displayed using dots with each one indicating a crime. Multiple instances can then be aggregated to display clus- ters. Different classes can be created to show high crime locations.

This technique is useful in identifying high crime locations. Specifically, the aggregation of crime events is particularly suitable when trying to identify high crime locations. Because this tech- nique is location specific, conducting temporal or crime compari- sons is not visually suitable on a single map. Rather, two maps need to be placed side by side in order to compare things such as crime events by time of day, crime type or over time. Additionally, as the density of events at a particular location increases, these dot maps become difficult to interpret. If one dot represents each event a high volume location becomes saturated with dots quickly. This issue can be resolved to some extent with the use of graduated dots (larger dot for a greater number of points). Finally, another signif- icant concern with this technique, especially when used for public distribution, is individual privacy (Kounadi, Bowers and Leitner, 2015). Privacy concerns arise in areas where there are fewer

Fig. 2. Street profile technique.

Fig. 3. Kernel density comparative visualization 250 m-100 m-50 m Rasters.

V. Spicer et al. / Applied Geography 69 (2016) 65e74 69

crimes and the marked crime location can potentially identify the victim.

4.3. Street segment crime density

This third analysis technique is not as commonly used in crime analysis, but has become common within the crime and place literature - see Weisburd (2015) for a recent review and discussion of this literature. In Fig. 5, the crime events are aggregated to the street segments and, like the aggregate count to address, crime events on street segments can be further aggregated and placed into defined classes. Research that investigated the trajectories of street segments over time has labeled them in the various

permutations of low, medium, and high-crime as well as stable, increasing, and decreasing (Curman et al., 2015; Weisburd et al., 2012).

This visualization technique is very useful in order to identify high crime street segments (Curman et al., 2015; Weisburd et al., 2012). These high crime places could be further analyzed in order to determine the environmental dynamics in these locations. However, like the previous technique, this one also has comparative limitations. In order to visually compare street segments for such things as night and day crime, longitudinal analysis or crime type comparison, two or more maps would need to be compared.

Fig. 4. Aggregate count to address.

V. Spicer et al. / Applied Geography 69 (2016) 65e7470

4.4. Street profile

Unlike the three previous techniques, the street profile tech- nique is not presented on a map, but rather on a graph. This sim- plifies the visualization and therefore allows for comparative analysis on a single chart. In Fig. 6, crime events are displayed using a line graph and this single line shows how crime fluctuates along a roadway. In this first example, three separate years are compared (2003e2004 e 2005). This longitudinal analysis shows how crime is increasing at regional shopping centres and becomes an obvious crime attractor. Table 1 accompanies the map to show the actual percentage increase as well as the raw numbers.

There are several benefits to this technique. First, the graph is simple to read even for non-subject matter experts. Second, the graph describes the crime dynamic well e is crime going up or down? Third, the graph shows the variation of crime on roadways and helps clearly define high crime places.

Other comparisons can also be completed. In Fig. 7, crime by day and night are compared. Clearly, more crime occurs during the day, which is congruent with this particular high crime location e a regional shopping centre.

In Fig. 8, crimes are compared by type. Again, this further clar- ifies the problem with property crime prevailing, also explained by the type of location.

4.5. Transect analysis

The transect analysis is a means to describe in a static manner the dynamics of crime directionality that is exposed using the street profile analysis. The street profile graph reveals a significant crime surge at the mall with two other lower surges at the nearby intersecting streets (Willingdon Ave and Royal Oak Ave). A further analysis of these intersecting streets using the transect methodol- ogy shows the prominent directionality of crime along Kingsway. In Fig. 9, 50 m buffers are used to demonstrate directionality with the line transect at intersections. There exists an eastward pull on Kingsway between the intersections of Willingdon Ave and Royal Oak Ave. The directional aspect of crime dissipates quite rapidly towards the north and south of Kingsway. As well, the crime den- sity at these intersections is highly varied with a higher density on Kingsway. This contrast is of particular note at Kingsway and Willingdon where crime density is both at the highest crime den- sity category in the Kingsway buffers and second lowest density in the Willingdon buffer.

5. Conclusion

In this study, we explore a new technique for understanding crime in small places within the urban domain. This mapping technique utilizes a graph approach that can be applied to major roadways in urban areas. While this technique is applied to

Fig. 5. Street segment crime density.

Fig. 6. Street profile crime density.

V. Spicer et al. / Applied Geography 69 (2016) 65e74 71

reported crime on a roadway, this method would also be useful in other types of analysis pertaining to roadways such as traffic analysis.

In this study, Kingsway is a major pathway for vehicles, a light-

rail mass transit system (Skytrain) line that runs parallel, a bike path that also runs parallel, and pedestrians who attend the area for business, shopping and entertainment. The study area contains very prominent activity nodes such as the Skytrain station and the

Table 1 Percentage crime by year.

2003 2004 2005

Crime counts 6524 8465 9526 Increase rate (%) e 29.8% 13.7%

Fig. 7. Street profile: night and day comparison.

Fig. 8. Street profile: crime type comparison.

V. Spicer et al. / Applied Geography 69 (2016) 65e7472

largest shopping mall in British Columbia. There are interesting temporal variations that are revealed using this technique that allow practitioners and policy makers to better understand the crime dynamics of major roadways.

This graph approach utilized to display major roadways allows for numerous comparisons that can help further understand the dynamics of these places. In particular, this visualization is easy to interpret, making it a good tool for describing crime problems to policy makers and civic personnel. The most common spatial vi- sualizations are displayed on maps such as kernel density and aggregate address counts and these are not as visually simple as the

street profile. Comparative analyses using maps requires multiple maps, whereas the street profile technique allows for comparisons on a single graph. Moreover, because of their calculations, these other methods are prone to false inferences regarding the location they represent, particularly kernel density. However, the street

profile method handles temporal and longitudinal analysis very well and can help expose the growing nature of a crime generator.

Analyzing major roadways is a means to better understand crime distribution and, thus, allocate resources. In certain in- stances, major roadways can be densely distributed crime areas where crime does not bleed significantly past these areas. This ef- fect is shown when looking at the transecting streets. In this study, the streets that cross Kingsway do not experience the same crime surge as there is along Kingsway. Enforcement would likely be more effective if it mimicked this crime pattern with concentrated enforcement along the roadway and targeted crime prevention

Fig. 9. Line-transect: density buffer analysis.

V. Spicer et al. / Applied Geography 69 (2016) 65e74 73

with the businesses and multi-dwelling residences in that area. Future research into this visualization technique will utilize data

from other major cities in order to further define the dynamics that form these places. The street profile method will be used to look at and compare different values. In this study, only crime is used to form the street profile. However, future research will compare crime to other civic data such as transportation and pedestrian traffic flow. This will allow for a more comprehensive under- standing of crime in the urban domain.

References

Bailey, T. C., & Gatrell, A. C. (1995). Interactive spatial data analysis. Harlow, UK: Prentice Hall.

Brantingham, P. J., & Brantingham, P. L. (1981). Notes on the geometry of crime. In P. J. Brantingham, & P. L. Brantingham (Eds.), Environmental criminolgy, prospect heights. Waveland Press.

Brantingham, P. L., & Brantingham, P. J. (1984). Patterns of crime. New York, NY: Macmillan.

Brantingham, P. L., & Brantingham, P. J. (1993a). Environment, routine and situation: toward a pattern theory of crime. Advances in Criminological Theory, 5, 259e294.

Brantingham, P. L., & Brantingham, P. J. (1993b). Nodes, paths and edges: consid- erations on the complexity of crime and the physical environment. Journal of Environmental Psychology, 13(1), 3e28.

Brantingham, P. L., & Brantingham, P. J. (1995). Criminality of place: crime gener- ators and crime attractors. European Journal of Criminal Policy and Research, 3(3), 5e26.

Brantingham, P. L., & Brantingham, P. J. (2015). Understanding crime with compu- tational topology. In Martin A. Andresen, & G. Farrell (Eds.), The criminal act: the role and influence of routine activity theory. New York: Palgrave Macmillan.

Chainey, S., & Ratcliffe, J. (2005). GIS and crime mapping. Hoboken, NJ: Wiley. Chainey, S., Tompson, & Uhlig, S. (2008). The utility of hotspot mapping for pre-

dicting spatial patterns of crime. Security Journal, 21, 4e28.

Curman, A. S. N., Andresen, M. A., & Brantingham, P. J. (2015). Crime and place: a longitudinal examination of street segment patterns in Vancouver, BC. Journal of Quantitative Criminology, 31(1), 127e147.

Davies, T., & Johnson, S. D. (2015). Examining the relationship between road structure and burglary risk via quantitative network analysis. Journal of Quan- titative Criminology, 31, 481e507.

Eck, J., & Weisburd, D. (Eds.). (1995). Crime and place, crime prevention studies (Vol. 4, pp. 1e34).

Frank, R., Andresen, M., & Brantingham, P. (2012). Criminal directionality and the structure of urban form. Journal of Environmental Psychology, 32(1), 37e42.

Frank, R., Andresen, M. A., Cheng, C., & Brantingham, P. (2011). Finding criminal attractors based on offenders' directionality of crimes. In Intelligence and se- curity informatics conference (EISIC), 2011 European (pp. 86e93).

Groff, E. R., Weisburd, D., & Yang, S.-M. (2010). Is it important to examine crime trends at a local “micro” level? A longitudinal analysis of street to street vari- ability in crime trajectories. Journal of Quantitative Criminology, 26(1), 7e32.

Herrman, C. R. (2013). Street-level spatiotemporal crime analysis: examples from Bronx County, NY (2006-2010). In M. Leitner (Ed.), Crime modeling and mapping using geospatial technologies. Dordrecht: Springer.

Johnson, S., & Summers, L. (2015). Testing ecological theories of offender spatial decision making using a discrete choice model. Crime and Delinquency, 61(3), 454e480.

Knapp, J. (2013). Safety and urban design e the role of CPTED in the design process. Safer Communities, 12(4), 176e184.

Kounadi, O., Bowers, K., & Leitner, M. (2015). Crime mapping on-line: public perception on privacy issues. European Journal on Criminal Policy and Research, 21(1), 167e190.

Loukaitis-Sideris, A. (2014). Fear and safety in transit environments from the women's perspective. Security Journal, 27(2), 242e256.

Manly, B., & Navarro Alberto, J. (2015). Introduction to ecological sampling. Boca Raton, Florida: CRC Press.

Newton, A. D. (2004). Crime on public transport: ‘Static’ and ‘non-static’ (moving) crime event. Western Criminology Review, 5(3), 25e42.

Ratcliff, J. (2006). A temporal constraint theory to explain opportunity-based offending patterns. Journal of Research in Crime and Delinquency, 43(3), 261e291.

Rengert, G. F., & Wasilchick, J. (1985). Suburban burglary: a time and place for

V. Spicer et al. / Applied Geography 69 (2016) 65e7474

everything. Springfield, IL: Charles C. Thomas. Skogan, W. (2015). Disorder and decline: the state of research. Journal of Research in

Crime and Delinquency, 52(4), 464e485. Smith, P., Phillips, T. L., & King, R. D. (2010). Incivility: the rude stranger in everyday

life. Cambridge: Cambridge University Press. Song, J., Frank, R., Brantingham, P., & LeBeau, J. (2012). Visualizing the spatial

movement patterns of offenders. In SigSpatial 2012 Proceedings of the 20th in- ternational conference on advances in geographic information systems.

Song, J., Spicer, V., Brantingham, P., & Frank, R. (2013). Crime ridges: exploring the relationship between crime attractors and offender movement. In Conference proceedings e European Intelligence and Security Informatics Conference EISIC (2013).

Spicer, V. (2012). The geometry of fear: an environmental perspective on fear and the

perception of crime. PhD Dissertation. Burnaby BC: Simon Fraser University. Van Daele, S., & Bernasco, W. (2012). Exploring directional consistency in offending:

the case of residential burglary in the Hague. Journal of Investigative Psychology and Offender Profiling, 135e148.

Vandeviver, C., Van Daele, S., & Vander Beken, T. (2015). What makes long crime trips worth undertaking? Balancing costs and benefits in burglars' journey to crime. British Journal of Criminology, 55(2), 399e420.

Weisburd, D. (2015). The law of crime concentration and the criminology of place. Criminology, 53(2), 133e157.

Weisburd, D., Groff, E. R., & Yang, S. M. (2012). The criminology of place: Street seg- ments and our understanding of the crime problem. New York, NY: Oxford Uni- versity Press.