Keywords

1 Introduction

The topic of urban security has become highly relevant in the urban agendas of cities and metropolitan areas. One of the main requirements for well-functioning cities and their sustainability is that they have secure streets and public spaces [4, 5] as confirmed by the Sustainable Development Goals (SDGs) of the 2030 Agenda, in particular SDG 11 ‒ Make cities and human settlements inclusive, safe, resilient and sustainable and target 11.7 – Access to safe and inclusive public space [1].

Although security is recognised as a fundamental human right, it is a resource as essential as it is rare and, for this reason, must be pursued and protected. The daily occurrence of predatory crimes – assaults, thefts, robberies, muggings, pickpocketing and other violent acts ‒ or episodes of incivility ‒ environmental and social ‒ leads to an increase in insecurity, both real and perceived by citizens, and fear of crime [6, 7]. Very often insecurity is associated with fear of those who are different. This situation leads to negatively labeling entire social groups and parts of city generating a closure to the outside world, either out of defense or spatial claim. Widespread fear becomes a real social problem with actual and measurable impacts on the city’ s functioning and the economy of the whole urban structure.

The achievement of security conditions can only be realized through good governance, planning and urban management, based on an integrated approach [8]. This is confirmed by the contents of the United Nations Guidelines for the Prevention of Crime [9] and as witnessed by some important International and European initiatives such as the now completed United Nations Safer Cities Programme [10], the COST Action TU1203 ‒ Crime Prevention through Urban Design & Planning (CP-UDP), the European project Cutting Crime Impact (CCI), the establishment in 2019 of the Security in public spaces partnership of the Urban Agenda for the European Union [11].

«Since ancient times, the need for personal security has conditioned the way cities are built. Ditches have been dug, walls erected, in an attempt to physically separate the city with its citizens from the outside world, identified as a source of danger and threats to urban life. The difference from the past is that in the modern and postmodern city, danger is no longer outside. Security threats are inside the city» [12], p. 124]. This calls for a review of how this issue has been, and is being, addressed.

Both internationally and nationally, for a long time the problem of urban crime was addressed only through two approaches: traditional (law and order) and social. Since the second half of the last century, a third approach has been introduced: the environmental approach. This one is known internationally as Crime Prevention Through Environmental Design (CPTED) and is the field of action of urban planning on the issue. CPTED is a preventive practice whose attention is focused on the urban environment and on its potential to hinder illegal behavior and transmit confidence to population. This approach, initially only focused on the physical characteristics of spaces and safeguarding of the most vulnerable areas, has been enriched with new elements over time, opening up to the problem's social dimension as well. In fact, it’s possible to identify three generations of this one [2]. Since its usefulness was formally recognized in 2001 in Sweden at the Towards a knowledge-based strategy to prevent crime conference, its basic principles have inspired the creation of standards and technical reports. These include the CEN/TR 14383-2: 2007 – Part 2, and the now international standard: ISO 22341:2021.

Unlike in the past, when these approaches were considered alternatives to each other, there is now recognition of the importance of integrating them in order to achieve concrete and lasting results. It is clear that the objective of making the city safer cannot be pursued exclusively through repressive or control actions or sporadic social interventions. The exclusive use of such actions, the overconfidence placed in them and the political exploitation of the problem only further increase the feeling of insecurity. The importance of urban security requires a change of perspective and more careful reflection, aimed at responding concretely to citizens’ growing demand for security. This need is, today more than ever, indispensable, also in combination with the need for a rethinking of the traditional paradigms of urbanism, by virtue of adapting them to the needs of social distancing imposed by the recent, and still not overcome, Covid-19 pandemic [13]. The problem of urban crime is linked to the city' s physical and functional structure and its functioning. So, it is necessary to address the problem through an integrated approach in which socio-economic actions work together with physical and functional interventions in the built environment, in accordance with the environmental crime prevention strategies [2, 14,15,16,17].

The issue's complexity requires a systematic action on the territory which, from an urban planning point of view, should result in the introduction of explicit guidelines for crime risk prevention in the municipal urban Plan.

This article explores an ongoing research study in which the topic of urban security is addressed in terms of risk. More precisely, crime risk is linked to three risk factor-maps: crime hazard (Hc), crime vulnerability (Vc), crime exposure (Ec) and described through a spatially explicit composite crime risk index, IRc [2, 3]. In this work the attention is focused on Vc and on the possibility to reduce its critical levels by defining and monitoring of specific urban planning actions to be introduced in the municipal urban Plan.

2 Materials and Methods

2.1 Disaster Risk

Disaster risk represents the likelihood that a given physical phenomenon will produce human or property loss or economic damage by investing with a given intensity, and in a given interval of time, a territory with a specific level of vulnerability. It is the product of three risk factors: hazard (H), vulnerability (V) and exposure (E). The risk fundamental equation is as follows:

$${\text{R}} = {\text{ H x V x E}} = {\text{ H x D}}$$
(1)

in which D represents the potential impact and is obtained as a product of V and E.

The risk analysis process and the results representation are linked to the analysis aim, the availability of data and the way in which uncertainties are managed. The European Commission recommends five types of analysis [18, 19]:

  • qualitative,

  • semi-quantitative,

  • semi-quantitative based on composite indicators methodology,

  • quantitative deterministic,

  • quantitative probabilistic.

Among this, in the semi-quantitative approach based on composite indicators methodology the analysis results are represented by using risk matrices and risk maps. The risk matrices, in particular, allow to compare risks linked to different hazards with specific likelihood [19] and represent effective tools to define the risk acceptability (Fig. 1a). These are two-dimensional tables that combine event likelihood (H) and impact (D) and can be set up symmetrical or weighted (Fig. 1b) [18].

In health and social risk assessment, reference is made to five risk classes [20]: very low risk (R1), low risk (R2), moderate risk (R3), high risk (R4), very high risk (R5). These are constructed through the risk matrix and are linked to the five intensity levels that can be assigned to each variable on which the risk is dependent.

Fig. 1.
figure 1

Source: Authors’ Elaboration.

RM method: a) Risk matrix template. Source: [19]; b, c) Matrix for risk-ranking in five classes (R1, R2, R3, R4, R5).

Full size image

The assignment of the five risk classes to the analysis results exploits the matrix obtained from the combination of H and D. The intensity levels associated with the two factors are combined in the matrix cells according to the logical Boolean operator AND and identify a specific intensity class of the phenomenon under study. For example, in the cell 11 are combined hazard map's intensity level 1 and impact map’s intensity level 1. This combination is symbolically referred as 11 and correspond to R1 risk class (Fig. 1c). The method just described is identified in following as RM (Risk Matrix) method.

2.2 Spatial Analysis Techniques

2.2.1 Kernel Density Estimation

Kernel Density Estimation (KDE) is a spatial analysis technique which allows to estimate the intensity variation of a given phenomenon in the study area. It is part of Point Pattern Analysis. Starting from a point pattern, KDE associates a density value to each point p of space and then calculates the number of events contained within a given region of the space (kernel). The last is centered at the estimation point. The density at point L is measured by using the following expression [21, 22]:

$$\lambda (L) = \sum\nolimits_{i = 1}^{n} {\frac{1}{{\tau^{2} }}k\left( {\frac{{L - L_{i} }}{\tau }} \right)}$$
(2)

in which:

  • λ(L) = estimated density at point L,

  • Li = generic event,

  • k(…) = kernel which is defined by a probability density function [3, 21,22,23],

  • τ = bandwidth.

The analysis result is influenced by the type of distribution, the bandwidth and the width of reference grid cell used to estimate the density values.

2.2.2 Space Syntax Analysis

Space Syntax Analysis (SSA) allows to measure accessibility of urban space by considering the spatial organization of cities as a factor affecting the pedestrian movement within it. The configurational approach considers the visual perspective connecting two points as the unit of measurement and a guide of movement within the urban grid. This one can be modeled in different ways based on the specific operational technique chosen: Linear Analysis (LA), which is divided into Axial Analysis [24], Angular Analysis and Angular Segment Analysis [25], Fragment Angular Analysis [26]; Visibility Graph Analysis (VGA), which can be normal or angular [27]; Mark Point Parameter Analysis (MaPPA) [28].

The Angular Segment Analysis (ASA) considers as basic element the segments obtained by splitting axial lines [24] at their intersection points. Starting from the segment map, this operational technique identifies the shortest route in a specific trip [25] by evaluating the sum of the rotation angles from the starter segment to any other within the system.

Among the configurational parameters, the most important is the integration index [29,30,31]. The last is specific for each segment and identifies the value of mean depth of this one in relation to all others in the segment map. It is associated with to movement [30]. A space is integrated if its depth is low, on the contrary it is segregated in correspondence of high depth values. The equation to evaluate the length-weighted mean angular depth of the segments (Cɵl) is as follows [25]:

$$C_{\theta }^{l} (x) = \frac{{\sum\nolimits_{i = 1}^{n} {d(x,i)l(i)} }}{{\sum\nolimits_{i = 1}^{n} {l(i)} }}.$$
(3)

in which:

  • \(C_{\theta }^{l}\) = weighted mean angular depth,

  • n = number of x segments in the graph,

  • dɵ = total angular depth,

  • l = length of the x segment.

Another important configurational measure is the betweenness (or choice) index, which represents through movement [30]. It evaluates the potential of a given segment to be chosen as a route by or to all other segments during the travel within the urban grid. The high potential is associated to high values of the index. The equation to evaluate the weighted angular choice (Bɵl) is as follows [25]:

$$B_{\theta }^{l} (x) = \sum\nolimits_{i = 1}^{n} {\sum\nolimits_{j = 1}^{n} {\sigma^{l} } } (i,x,j)\;{\text{such}}\;{\text{that}}\;{\text{i}} \ne {\text{j}}$$
(4)

in which:

  • \({\text{B}}_{\theta }^{1}\) = weighted angular choice,

  • n = number of i, x, j segments in the graph,

  • l = length of the x (i or j) segment.

Finally, connectivity identifies the number of segments directly accessible starting from the examined one. It represents the number of perspective openings of each segment to other convex spaces and its availability of movement to other segments [2].

3 Methodology

In the research mentioned above, crime risk is modelled according to the European Commission's recommendations on risk assessment and mapping and, more precisely, follows a semi-quantitative approach based on composite indices [18,19,20].

The IRc index is structured according to the territorial risk paradigm and spatialized in a crime risk map [2, 3]. It combines the main aspects that contribute to the characterisation of risk scenarios through three risk factor-maps (Hc, Vc, Ec), described by appropriate spatial indicators (Fig. 2), as follow:

$${\mathrm{IR}}_{\mathrm{c}}={\mathrm{H}}_{\mathrm{c}}\mathrm{ x }{\mathrm{V}}_{\mathrm{c}}\mathrm{ x }{\mathrm{E}}_{\mathrm{c}}= {\mathrm{H}}_{\mathrm{c}}\mathrm{ x }{\mathrm{D}}_{\mathrm{c}}$$
(5)

in which Dc represents the crime impact obtained by combining the two factor-maps: Vc and Ec. The IRc spatialization is performed by implementing density, proximity and spatial configuration analyses and through the use of appropriate map algebra operations. In detail the techniques and the operations used are respectively: KDE [22], Euclidean Distance (ED) function [32], LA [24], classification and overlay mapping between raster [33]. The crime risk map is structured in five risk classes: very low risk (R1), low risk (R2), moderate risk (R3), high risk (R4), very high risk (R5) [2, 3] which are assigned by using the RM method mentioned above (Sect. 2).

Fig. 2.
figure 2

Source: Authors' elaboration.

Methodological scheme of crime risk map construction.

Full size image

The approach used considers risk as a convolution of the three mentioned factor-maps and is based on the assumption that it can be reduced or eliminated when at least one of them is reduced or eliminated. Therefore, it is possible to examine each risk factor-map separately in order to outline the most suitable strategies to reduce its critical conditions and, consequently, reduce crime risk.

This study focuses in particular on Vc, in relation to which urban planning has full control. In the research, crime vulnerability is described as the ability of urban space – understood in terms of its physical and functional layout – to discourage or favour the occurrence of opportunity crimes. Vc is constructed as a combination of four spatial indicators: marginality (M), functional mix (Mf), proximity to decayed areas and buildings (Pro) and accessibility (Ac), each of which is a function of specific parameters [2, 3]. Among these, M and Pro were selected as baseline variables of Vc because the most marginal areas of city and the decay of areas and/or buildings are linked to a greater tendency to attract micro-criminality and incivility [14,15,16]. Therefore, both are considered directly proportional to vulnerability. Mf and Ac were selected as physical-functional characteristics through which it is possible to influence informal surveillance of spaces. In particular, the former is associated with more eyes on the street [5] as, if correctly calibrated, it can generate more vitality and continuous use of space. The second is a necessary condition, together with intervisibility and permeability, to ensure that spaces can be read and easily crossed by residents and non-residents alike [29,30,31]. For these reasons, both variables are considered inversely proportional to vulnerability.

Vc is structured in five classes: very low vulnerability (V1), low vulnerability (V2), medium vulnerability (V3), high vulnerability (V4), very high vulnerability (V5). The maximum values, which correspond to the highest concentrations of the four base maps (M, Mf, Pro, Ac), identify the most vulnerable areas in relation to which appropriate intervention strategies should be planned.

3.1 Security Actions for the Municipal Urban Plan

The variables that contribute to determining crime risk conditions are numerous. Its critical levels may vary in space and time as a result of social, cultural and economic dynamics, but also as a consequence of urban planning actions. The criteria for planning, design and construction of cities can reduce crime risk or increase it, if not properly defined and assessed [14, 34]. In order to reduce it, urban security must be taken into account from the earliest stages of planning, by addressing its aspects in all urban planning tools, be they structural plans, operational plans, sectoral plans, urban regeneration programmes, major equipment projects, etc. The action, carried out following a multi-scalar approach, allows to simultaneously act on all aspects affecting the vulnerability conditions in order to ensure safer urban spaces.

According to this observation, the research proposes specific actions to be introduced in the municipal urban Plan to promote a systematic adaptation of spaces to security criteria [2, 3], acting both directly and indirectly on Vc (Table 1). These were defined as a result of a complex operation of systematisation and synthesis of data obtained from state of art reconstruction on the topic, carried out by analysing manuals, technical-scientific literature, twenty-five guidelines and twenty-five case studies relating to the application of environmental crime prevention principles. Therefore, starting from the elements of physical and functional organisation that showed the greatest relevance and recurrence, actions were defined for the Structural Plan (SP), the Operational Plan (OP) and the Urban Regulation (UR). Specifically, the security actions outlined for the SP are oriented to affect informal surveillance of urban space by acting on the city’ vitality and on its time of use through suitable location choices of activities and land uses. The actions for the OP are aimed at strengthening the feeling of territoriality, i.e., the citizens’ sense of belonging to the area in which they live [16, 17], and the possibility to exercise both informal and formal surveillance. They concern the structure of urban spaces, green areas, car parks, roads etc. In the UR it is intended to introduce actions mainly aimed at guaranteeing visual permeability from public and private spaces to public spaces and, therefore, informal control, and at enhancing the comfort of places by reducing their propensity to decay. This is possible through several measures including: the correct definition of semi-public spaces, routes, etc.; the type of urban green spaces and the criteria for its correct distribution; a correct lighting, transparency and appropriate fencing etc.

Table 1. Security actions and their influence on Vc.
Full size table

Among the actions outlined, those proposed for the SP mainly affect M and Mf. The actions outlined for the OP affect the variables Mf, Pro and Ac. Some of the actions outlined for the UR directly influence Mf and Ac and indirectly Pro (Table 1). For example, guaranteeing the continuity of urban structure, of road layout and cycle-pedestrian network and the correct location of commercial activities and services ensures equal accessibility conditions to spaces, infrastructures and services in all areas of the city, avoiding the creation of marginal zones or enclave situations, directly affecting M and also Ac. Again, actions oriented to place public spaces in densely populated areas, activities for continuous and night use in strategic points, commercial activities and public services also in the ground floors of residential buildings directly influence Mf, increasing and diversifying land uses to guarantee a continuous use of spaces. In this sense, it can be seen that some functions are more effective than others in ensuring vitality. The activity generators, i.e., schools, universities, hotels, public buildings, shops and commercial centres, which, wisely placed in the territory are able to increase the number of eyes on the street but also to attract other functions. In addition, by ensuring a clear distinction between public, semi-public, semi-private and private spaces, made possible through symbolic boundaries (by diversifying materials, paving, introducing flower boxes, benches, etc.) it is possible to indirectly influence Pro. This action, in fact, is oriented to increase territoriality based on the principle that people tend to defend and take care only of what they consider to be their own. The care and maintenance of spaces reduces their degradation.

3.2 Crime Vulnerability Scenarios for Monitoring Urban Planning Actions

The changes in physical and functional organisation of urban spaces, as a result of urban planning choices, have impacts on Vc and therefore also on crime risk. Urban planning actions can reduce crime vulnerability but can also increase it, if not properly assessed, causing an increase in existing criticalities or triggering new ones.

In order to reduce Vc, it is necessary to start a systematic adaptation of spaces to security criteria, making urban planning choices aligned with the environmental crime prevention principles, applying specific urban planning actions and rules (Sect. 3.1) and monitoring their impact on the territory. In response to this need, the risk model outlined provides for the construction of crime vulnerability scenarios (Vc,i, with i = 1, …, n) [2] (Fig. 3). The construction of these makes it possible to construct n crime impact scenarios (Dc,i, with i = 1, …, n) and n crime risk scenarios (IRc,i, with i = 1, …, n) corresponding to them, according to the following expression:

$${\mathrm{IR}}_{\mathrm{c},\mathrm{i}}={\mathrm{H}}_{\mathrm{c}}\mathrm{ x }{\mathrm{V}}_{\mathrm{c},\mathrm{i}}\mathrm{ x }{\mathrm{E}}_{\mathrm{c}}= {\mathrm{H}}_{\mathrm{c}}\mathrm{ x }{\mathrm{D}}_{\mathrm{c},\mathrm{i}}$$
(6)
Fig. 3.
figure 3

Source: Authors’ elaboration.

Methodological scheme for constructing crime vulnerability and crime risk scenarios, and for integrating of the IRc in urban planning tools.

Full size image

These scenarios are important cognitive frameworks. If used in the ex ante and/or ex post phase of the municipal urban Plan formation, these ones allow to make informed choices, to outline an adequate framework of actions and rules, and to assess the need for corrections to them if the urban Plan is already in force. Through these scenarios it is possible, in fact, to highlight the criticalities in territory, to simulate the effect that urban planning choices may have on existing Vc levels (which may consist of a decrease or increase of these or in triggering new criticality); to monitor, over time, the results of environmental crime prevention strategies.

The model outlined allows to simulate, for example, the effect of actions on Mf such as refunctionalisation interventions; also, actions oriented to reduce M and Pro such as regeneration and recovery of degraded situations by changing the input variables associated with the different land use categories and modifying the number and surface extent of degraded areas. Also, it is possible to simulate future scenarios showing changes in vulnerability conditions linked to the implementation of forecasts of the urban planning tool in force in order to identify the need for corrections.

4 Application of the Model

4.1 Case Study

The model was applied to the city of Milan, which was chosen for the availability of data on crimes, urban planning tools and elements of urban structure (Fig. 4).

Fig. 4.
figure 4

Source: Authors' elaboration.

Territorial context of the study area; a) Municipal territory divided into 88 neighbourhoods (NIL); b) Crime risk map [2], c) Zoning, PGT-Milano 2030.

Full size image

The city, as of today, corresponds with the only statistically significant dataset that is available and can be used for research purposes in relation to the crimes under study. This was confirmed by the results of an investigation conducted on a selected sample of Italian cities, to whose Police Headquarters and Prefectures were sent a request for georeferenced data about crime occurred in their Municipality in a given time interval, and by subsequent research regarding the spread of crime mapping practices nationwide [2].

Starting from the crime risk map constructed for the city in a previous study [2, 3]. (Fig. 4b), the focus shifted from the urban to the detail scale. More precisely, the attention was focused on the Villapizzone-Cagnola-Boldinasco neighbourhood (Fig. 5). The latter is one of the 88 neighborhoods identified as Nuclei di Identità Locale (NIL) by the Piano di Governo del Territorio (PGT Milano-2030) which corresponds to the urban Plan of Milan [35]. The selected neighborhood is located in the northwest of the municipal territory, covers an area of 3.4 km2, and is inhabited by 22.062 families with a population density of 12.178 inhab/km2. The latter has a very high degree of urbanization, accounting for 92.6% of the total area. In detail, the prevailing land uses are divided as follows: 27.9% of the area coincides with built-up fabric, 24.1% with unbuilt-up fabric, 27.2% is for mobility and 21.4% is for green space. In PGT-Milan 2030 zoning (Fig. 5a), the neighborhood is characterized by the predominant presence of Ambiti di Rinnovamento urbano (ARU), which correspond to urban renewal areas, environmental regeneration areas, which are flanked by existing and newly planned urban green spaces. In addition, the presence of external historic cores and pedestrian-oriented spaces is found [35, 37].

The neighbourhood was selected for this study as it is characterized by high critical conditions in terms of crime risk and crime vulnerability (classes R3, R4, R5 and V3, V4, V5 respectively) and by a high concentration of urban renewal areas and areas of environmental regeneration, as mentioned above.

4.2 The Crime Risk Map (CRM) Model

The scenarios simulation is automated through the construction of a parametric model realised as a Model Builder through ArcGIS software [38, 39]: the Crime Risk Map (CRM) model [12]. The latter automates the procedure of crime risk map construction making it replicable in any context and in very few minutes. It is structured in four parts, the most complex of which is the one dedicated to Vc construction (Fig. 6). In fact, the last requires density analysis, proximity analysis, spatial configuration analysis and several steps of combination and assignment of criticality classes according to the RM method. The parametric model realised automates these operations through 9 functions and several parameters acting on 12 input variables.

Fig. 5.
figure 5

Source: Authors' elaboration starting from PGT-Milano 2030 charts. Please refer to the following link for the plan charts used: https://www.pgt.comune.milano.it/pgt-previgente/pgt-milano2030-approvato-05022020/piano-dei-servizi-approvato-05022020/norme-catalogo-e-nil-05022020 (last accessed: 20/07/2022).

Territorial context of the neighbourhood under study: a) areas planned for regeneration and transformation; b) existing services; c) the green and water system, d) public buildings and spaces; e) mobility and accessibility.

Full size image
Fig. 6.
figure 6

Source: Authors’ elaboration.

Scheme of the Model builder constructed in GIS: excerpts.

Full size image

4.3 Constructing Scenarios of Crime Vulnerability and Crime Risk

The application was carried out by simulating two reference scenarios using the CRM model [12]. Scenario 0 describes the current state of the territory. In detail, for the neighborhood under study, Vc,0 (Fig. 7b) is characterized by critical levels (V3, V4 and V5) at the northern and central parts of the urban space. This circumstance is most influenced by conditions of unbalanced functional mix and the presence of decaying areas. Significant critical situations are also found in terms of crime risk. Indeed, as can be seen in IRc,0 (Fig. 7d), over half of the neighborhood surface is affected by medium (R3), high (R4) and very high (R5) levels of risk.

Starting from this evidence, the urban fabric was examined in detail with the help of satellite maps and the Google Street View tool, in order to deepen factors behind the neighborhood's high vulnerability and collect all the necessary elements to formulate a specific intervention proposal.

The study revealed the presence of entire blocks characterized by decay of building fronts, also marked by the presence of graffiti, especially in areas close to the railway infrastructure. The presence of the latter also serves as a physical barrier of separation from the rest of the neighborhood to the northeast. Even in this area, marked by the presence of railroad and industrial areas, a low quality of urban space was found, with the presence of elements of decay and graffiti. Such circumstances contribute to attracting petty crime and encouraging incivility and vandalism. In addition, the unbalanced functional mix is not supportive of informal control by citizens or city users [5, 31].

Based on the results of this step and in accordance with the PGT-Milan 2030 forecasts [35], a specific intervention proposal was formulated oriented toward reducing vulnerability levels by acting directly on Mf and indirectly on Pro with reference to some selected areas (Fig. 7a). In detail, the proposed actions involve urban regeneration interventions aimed at the introduction of public activities and services in strategic locations and recovery of decaying conditions, preferring the use of durable materials and anti-graffiti paints (Table 1).

Therefore, a new scenario, identified as scenario 1, was constructed by simulating the change in crime vulnerability and crime risk conditions resulting from the implementation of the proposed interventions. As can be seen, Vc,1 (Fig. 7c) shows a significant reduction in criticality levels particularly in the northern zone where class V5 has almost completely disappeared. This decrease in vulnerability results in a decrease also in the crime risk classes. Indeed, IRc,1 (Fig. 7e) shows significant improvement in the north side of the neighborhood.

Fig. 7.
figure 7

Territorial context of the neighbourhood under study: a) areas of interest on satellite view; b) Crime vulnerability map – scenario 0 (Vc,0); c) Crime vulnerability map – scenario 1 (Vc,1); d) Crime risk map – scenario 0 (IRc,0); e) Crime risk map – scenario 1(IRc,1); Source: Authors’ elaboration.

Full size image

5 Discussion and Conclusions

The model's application to the case study has shown the potential of the latter. The maps of scenario 0 (Vc,0 and IRc,0) showed critical levels both of crime vulnerability (V3, V4, V5) and crime risk (R3, R4, R5) mainly concentrated in the central and nortwest areas of the neighborhood. As clarified earlier (Sect. 4), these criticality in term of vulnerability are mainly caused by the presence of widespread decay and an unbalanced functional mix.

The identification of these critical issues and the proposed intervention at the basis of the scenario 1 simulation (Vc,1 and IRc,1) highlighted the potential of urban interventions in mitigating existing levels of crime vulnerability and crime risk and preventing new criticality.

In relation to the case study, such scenarios can be integrated into the municipal urban Plan at an ex post stage, as the PGT-Milano 2030 is now in force. Their integration makes it possible to assess the conformity of urban planning choices to security criteria and suggest possible corrections to them.

The comparison of the crime risk map constructed for Milan with the contents of the PGT-Milan 2030 highlighted a lack of attention of the latter in relation to security criteria, as shown in a previous study [2, 30]. In fact, some of the interventions allowed by the Norme Tecniche di Attuazione (NTA) which coincide with the urban Plan’s normative regulations [36], are not aligned with environmental crime prevention criteria and could cause increased levels of crime vulnerability and crime risk, but also trigger new criticality. For example, in relation to the external historic cores [36, article 15] are planned regeneration interventions by allowing the realization of any urban function and the increase of private services, craft activities, neighbourhood shops and service activities. However, for areas already affected by critical conditions in relation to crime risk, especially if high (classes R4, R5) it is essential to ensure a suitable functional mix by balancing the amount of these services with other land uses, including residential [29]. With specific reference to the neighborhood under study the allowed interventions, if implemented without taking into account their impact on security, will result in an even greater imbalance in the distribution of functions that will potentially cause a worsening of vulnerability and risk from crime conditions (already currently critical). In addition, with reference to ARU, the NTA [36, article 22] allow for requalification of existing public spaces, affecting how they relate to private ones and promoting the creation of new local systems of collective spaces. However, such an incentive could negatively affect the levels of crime vulnerability and crime risk if the resulting interventions are located on the territory and designed without consideration of risk factors. In addition, the consultation of Piano dei Servizi (PS), which coincide with the Milan’s Plan for infrastructures and public facilities [37] showed that many of the neighborhoods within the central core of Milan's consolidated fabric are affected by forecasts that, in the medium and long term, will lead to changes in accessibility conditions. Again, no consideration is given to the impact this factor has on vulnerability and risk from crime [29, 31].

According to the observations made, the forecasts contained in PGT-Milan 2030 should be modified taking into account their impact in terms of urban security, as well as the other needs they are aimed at responding to, either by indicating corrective factors or by providing a detailed urban regulation.

In order to better understand the potential of Vc,i and IRc,i, scenarios (Sect. 3.2), consider, for example, the planning of public spaces and facilities for public use. These elements should ensure a balanced urban development. In Italy their planning is regulated as territorial facilities per inhabitant and no indication is given as to their location on the territory. However, this aspect is of fundamental importance from a security standpoint as such spaces and services if not carefully sized and located could transform into attractors of crime [15]. The model outlined, through the construction of scenarios (Vc,i and IRc,i) would allow for their calibration both in terms of size and location in relation to the territorial context under study and to its levels of crime vulnerability and crime risk.

In addition, Vc,i and IRc,i, scenarios could be also complemented by the implementation of appropriate participatory techniques, through which to explore the subjective dimension of risk, and to outline the most suitable strategies to address not only actual risk but also perceived one [6, 7].

From the results obtained, further research developments are outlined that will be conducted by adopting a multi-scalar approach. These include an in-depth study, at-currently underway, geared toward introducing new variables into the Ec factor-map and simulating dynamic scenarios of crime exposure and crime risk by using geo-social media data [40]. Further in-depth study will be oriented toward exploring how different crimes patterns are distributed within various types of urban fabrics, focusing in particular on social housing neighborhoods. In addition, another study will deep the non-linear relationship between functional mix and crime rate and will be oriented to identify the optimal density value even in relation to urban fabrics other than the one currently examined.

In conclusion, it is observed how the issue of urban security is of great importance for urban policies. Preventing and mitigating crime risk and increasing levels of personal security is an important requirement for creating sustainable cities and communities, as other social, economic and environmental actions [41,42,43] and as the mitigation of other kinds of risk [44,45,46].

Very often, especially in Italy, the contribution of urban planning to urban security is overlooked and underestimated. On the contrary, the urban planning choices, if not correctly assessed, can affect crime vulnerability causing an increase in existing criticalities or triggering new ones. This increases crime risk conditions.

City security must be pursued systematically, adopting an integrated approach in which all possible actions or strategies work in synergy. For this purpose, it is essential to flank and integrate socio-economic interventions and traditionally adopted control actions with suitable environmental crime prevention strategies.

The work presented here is a small part of a wide-ranging research project exploring the issue of urban security and the possible contribution that urban planning can make to it. In doing so, the research adopts a constructivist approach [47]. The objective of the outlined risk model is not to solve the urban crime problem but to provide the decision-maker with a tool that help define the most suitable strategies to tackle it, in relation to the context's characteristics.

This article focuses in particular on Vc, proposing a set of specific actions to be introduced in the municipal urban Plan aimed at reducing its critical levels. The effectiveness of such actions is linked to knowledge of the specific context of interest and the impacts they may have in the short and long term. For this reason, the model outlined provides for the construction of specific scenarios (Vc,i and IRc,i) that are important cognitive frameworks to be considered in the ex ante and/or ex post phase of the urban Plan's formation, in order to evaluate the urban planning choices and outline a suitable framework of actions and rules for increasing the city's security.

In conclusion, the knowledge of the current vulnerability and risk conditions and of their variation caused by physical and functional changes of urban spaces, allows to orient the urban planning choices towards crime risk prevention for a better organization of the city according to security criteria.