Police use-of-force hotspots: Mapping and analysing police use-of-force events in San Francisco

Abstract

This study examines the spatiotemporal distribution of police use-of-force (UoF) incidents reported by the San Francisco Police Department. Drawing on geocoded incident data, we analyse the concentration of UoF across space, time, and officers, and assess whether UoF hotspots correspond to areas with higher proportions of racial and ethnic minority residents. Findings reveal marked concentration: UoF incidents cluster in specific locations and time periods, and a disproportionate share of incidents involve a subset of officers. Multi-officer encounters were associated with a significantly elevated probability of suspect and officer injury. Notably, spatial analysis revealed no over-representation of UoF hotspots in predominantly minority neighbourhoods, suggesting that the geographic distribution of force is not systematically tied to the racial composition of the areas in which it occurs.

Keywords

use of force hotspots spatial and temporal patterns injuries to officers or suspects

Introduction

The use of force (UoF) is integral to police work, as apprehending suspects and disarming offenders may require physical coercion (Cowell et al., 2021; Longridge et al., 2023). However, when officers apply excessive or unwarranted force—particularly in incidents with racial dimensions—the consequences can include widespread public outcry and lasting damage to police legitimacy (Fryer, 2019; Ouellet et al., 2019; Smith et al., 2022). UoF encounters also carry risk of serious injury to both officers and suspects (Alpert and Dunham, 1998; Henriksen and Kruke, 2020). Understanding when, where, and under what circumstances force occurs is therefore essential—not only to reduce harm, but to preserve the institutional legitimacy upon which effective policing depends.

Research on police UoF has examined a range of explanatory factors, yet most yield inconsistent results in predicting the occurrence of force (Adams and Alpert, 2023; Bolger, 2015; Cojean et al., 2020; Cowell et al., 2021; Klahm and Tillyer, 2010). Some situational patterns are nonetheless evident. The likelihood of UoF is elevated during nighttime hours, particularly in late-night and early-morning periods (Henriksen and Kruke, 2020; Hickman et al., 2015; Piza et al., 2023). It also increases when additional officers are present at the scene (Bolger, 2015; Henriksen and Kruke, 2020) and when the encounter is officer-initiated (proactive policing) rather than dispatched in response to a call (Assaraf et al., 2024; Terrill et al., 2023). These findings, however, remain largely confined to the behavioural and situational context of individual encounters. What has received far less attention is the geographic setting in which force occurs—a limitation identified nearly 50 years ago that largely persists today (Sherman, 1980).

This gap is all the more striking given the parallel development of place-based approaches in policing. Hotspot policing, the most extensively developed ecological strategy, is premised on the well-established spatial concentration of crime (Weisburd, 2015), directing police resources toward the small number of places that account for a disproportionate share of offending (Braga et al., 2019a, 2019b; Weisburd et al., 2024). Beyond the spatial dimension, temporal analyses have further elucidated the geographic dynamics of crime activity (Ariel and Partridge, 2017; Piza and O’Hara, 2014; Santos and Santos, 2015). Aligning available police resources with the right places at the right times is what renders place-based strategies effective (Braga et al., 2014, 2019a, 2019b; Braga and Weisburd, 2022; Smith et al., 2024; Weisburd et al., 2024; Weisburd and Telep, 2014). Yet this body of research has concentrated almost exclusively on reported crime (Ariel et al., 2016b, 2024; Gill et al., 2017; Schnell and McManus, 2022), calls for service (Ariel et al., 2016b; Braga and Bond, 2008), index offences (Braga et al., 2011), and arrests (MacDonald et al., 2016; Santos and Santos, 2016). Police UoF—despite its operational and legitimacy implications—has remained at the margins of place-based inquiry.

The result is two largely disconnected literatures: one examining the correlates of UoF without attending to place, and another examining place-based crime dynamics without attending to force. Although research on the micro-geographic and temporal concentration of police UoF has expanded in recent decades, important gaps remain. This study responds to calls to examine force concentration independently of crime events (Sorg et al., 2021), providing both geographic and temporal analysis of police UoF at the micro-level (census block) as reported by the San Francisco Police Department (SFPD). It further investigates encounter characteristics associated with officer or suspect injury and hands-off force. By integrating spatiotemporal analysis with injury prediction, this study bridges these two bodies of scholarship and offers a more comprehensive account of UoF dynamics.

Identifying concentrations of police UoF can inform deployment strategies by highlighting areas with elevated incident frequencies. This knowledge can guide targeted intervention and resource allocation. Moreover, understanding how and where UoF incidents cluster provides a foundation for analysing injury predictors: locations and times with elevated UoF frequencies may pose greater risks of harm to both officers and suspects. By directing efforts toward places where UoF is most prevalent, agencies may reduce incident rates in these hotspots and, in turn, mitigate associated injury risks—potentially improving citywide UoF outcomes (Weisburd et al., 2017). Such reductions could enhance public trust and bolster perceptions of police legitimacy (Ouellet et al., 2019).

Literature review

Use of force

The use of force in police-public encounters has been a focal point of research in policing scholarship. This sustained interest reflects the commitment of practitioners and policymakers to understanding how law enforcement agencies exercise their authority and to identifying effective means of regulating this power (Ariel et al., 2015; Assaraf et al., 2024, 2026). Although definitional boundaries remain contested, UoF is generally understood as the direct or indirect application of physical coercion by a police officer toward a member of the public (Hine et al., 2018) or as “effort required by police to compel compliance by an unwilling subject” (International Association of the Chiefs of Police, 2001). Officers may resort to force when there is an immediate threat to public or officer safety, or when an encounter otherwise necessitates physical intervention (Longridge et al., 2023). The public broadly expects police to use force when necessary and proportionate to achieve legitimate objectives (Alpert and Smith, 1994). In practice, however, some officers apply excessive force that is unwarranted given the circumstances, or fail to respond proportionately to tense encounters. Conversely, some individuals act verbally or physically aggressively toward officers. These dynamics can escalate to forceful arrests and the deployment of police force in response to civilian behaviour (Ariel et al., 2016a).

Although UoF is an extraordinarily rare event, its consequences extend far beyond the individual encounter (Hickman et al., 2015). Force incidents can erode public trust in the police, undermine perceptions of legitimacy, reduce willingness to cooperate, and impair agencies’ capacity to perform their duties effectively (Adams and Alpert, 2023; Ouellet et al., 2019). Excessive force can also damage institutional reputation (Tuch and Weitzer, 1997) and, more immediately, result in injuries or fatalities involving officers or suspects (Alpert and Dunham, 1998; Henriksen and Kruke, 2020).

Research on police UoF has concentrated on officer characteristics (Brandl and Stroshine, 2013; Jain et al., 2022), suspect characteristics (Smith et al., 2022), and encounter-level factors (Bolger, 2015; Tillyer, 2022). Prior work demonstrates that situational and behavioural variables play a central role in shaping UoF outcomes. Specifically, encounters involving more serious offences are 1.46 times more likely to result in force; encounters with resistant suspects are 2.97 times more likely to involve force; and suspects who are arrested are 4.34 times more likely to experience force (see Bolger 2015 meta-analysis). A key variable in suspect-focused research is race—both the suspect’s race and the neighbourhood’s ethnic composition. Studies consistently show that minority groups, and communities with high concentrations of minority residents, face a significantly elevated risk of UoF, rather than being randomly distributed across the city (Lautenschlager and Omori, 2019; Lersch et al., 2008; Smith et al., 2022). Additionally, lethal UoF by police—incidents resulting in civilian fatalities—occurs more frequently in neighbourhoods with higher percentages of Hispanic and Black residents (Durán and Loza, 2017; Holmes et al., 2019).

Alongside suspect-related factors, encounter characteristics provide crucial context for understanding police UoF. These include day of the week, time of day, whether the encounter was officer-initiated or dispatched, whether an arrest was made, the presence of additional officers or civilians, and whether the encounter occurred in a public or private setting (Bolger, 2015; Cojean et al., 2020; Henriksen and Kruke, 2020; Klahm and Tillyer, 2010; Tillyer, 2022). Research indicates that the likelihood of UoF is elevated during dusk, late-night, and early-morning hours (Henriksen and Kruke, 2020; Hickman et al., 2015; Piza et al., 2023); when additional officers are present (Bolger, 2015; Henriksen and Kruke, 2020); and when the encounter is officer-initiated (proactive policing) rather than dispatched in response to an emergency call (reactive policing; Assaraf et al., 2024; Terrill et al., 2023). Recently, Sorg et al. (2021) demonstrated that police UoF incidents concentrate at a small percentage of subway stations in Philadelphia—a finding consistent with the “law of crime concentration,” which holds that crime and related incidents cluster in specific locations.

Law of crime concentration and hotspots

A growing body of literature has demonstrated that crime is not evenly distributed across urban landscapes but instead concentrates in a small number of geographic units that account for a disproportionate share of criminal incidents (Ariel and Partridge, 2017; Braga et al., 2019a, 2019b; Sherman et al., 1989; Sherman and Weisburd, 1995; Smith et al., 2024; Weisburd, 2015, 2018; Weisburd et al., 2009a, 2012, 2024). The consistency of these findings has led Weisburd and colleagues to characterise this pattern as the “law of concentration of crime in place” (Weisburd, 2015; Weisburd et al., 2012). Today, the focus on geography, terrain, and spatial mapping in the study of crime patterns is more firmly established than ever as a foundational component of criminological theory (Sampson, 2013; Short et al., 2010; Weisburd et al., 2012; Wortley and Mazerolle, 2008). Nearly all human activity is tied to place, and crime is no exception: offending must occur somewhere within the spatiotemporal spectrum. This “crime and place” perspective has signalled to criminologists that analysis should concentrate at the micro level—commonly operationalised as “hotspots of crime” (Sherman et al., 1989).

Beyond spatial concentration, research has revealed that crime hotspots also exhibit temporal variation, with elevated activity during particular hours, days, and months (Johnson et al., 2008; Ratcliffe, 2004; Townsley, 2008). Studies have further examined hotspot stability, finding that micro-geographic crime concentrations often display significant persistence over time (Braga et al., 2011; Groff et al., 2010; Schnell and McManus, 2022; Weisburd, 2015; Weisburd et al., 2009b, 2012). For example, Weisburd et al. (2004) demonstrated that crime-prone street segments in Seattle maintained their hotspot status for 14 years—a stability attributed to the continued availability of criminal opportunities at these locations (Kumar et al., 2024; Weinborn et al., 2017). However, other research indicates that some hotspots are “hot” only during specific periods (Lersch, 2022; Santos and Santos, 2015). A recent study found that hotspots exhibited temporal instability, with locations shifting in and out of hotspot classification over time (Lersch, 2022). These mixed findings suggest the presence of both “chronic” and “temporary” hotspots (Gorr and Lee, 2012). Understanding the temporal and spatial dynamics of hotspots is therefore crucial for developing effective crime prevention strategies.

The present study

Examinations of the geographic distribution of policing activity and its outcomes remain scarce. To our knowledge, aside from Sorg et al. (2021), no existing research has explicitly linked hotspot frameworks to police UoF. This study responds to calls for further investigation of UoF concentration, offering both geographic and temporal analysis of police UoF at the micro level (census block), alongside examination of encounter characteristics associated with officer or suspect injury and hands-off force. Our two central questions are: (1) How concentrated are UoF incidents across time, place, and officers? (2) What encounter characteristics elevate the risk of injury and hands-off UoF during police encounters? To address these questions, we analyse the geographic and temporal distribution of UoF incidents reported by the SFPD.

This study makes two key contributions. First, we provide a geographic exploration at the micro level (census blocks), including spatial concentration analysis and precise mapping of UoF clusters. We also examine the concentration of force among officers and suspects, distinguishing between spatial and agent-level concentration. Second, given inconsistent findings regarding whether hotspots are “hot” only during specific periods (Lersch, 2022; Santos and Santos, 2015) or remain stable over time (Weisburd et al., 2004, 2009b), we accord equal attention to the temporal dimension alongside the spatial.

Methods

Study area

San Francisco is located in Northern California and has a population of 808,988 (United States Census Bureau, 2023). The city consistently ranks among the highest in the nation for per capita income, with a reported per capita personal income of $155,819 (FRED, 2023). The most prevalent crime types are larceny-theft, burglary, and motor vehicle theft. The San Francisco Police Department (SFPD) operates 10 district police stations across the city (SFPD, 2024b) and employed 1203 patrol officers in 2021 (SFPD, 2022a).

The most appropriate unit of analysis for micro-geographic research in this context is the census block.¹ Census blocks differ slightly from street blocks: whereas a street block is typically defined as a single block face delimited by four intersections, census blocks can encompass larger areas. In San Francisco, most census blocks coincide with street blocks, except in coastal areas and less residentially developed zones. At this level of analysis, we investigate whether police UoF concentrations align with the law of crime concentration and the bandwidth established by Weisburd (2015). We further test the influence of officer proactivity and multi-officer incidents on the prevalence of police UoF. Specifically, we ask: where does UoF concentrate, and what factors are associated with a lower risk of police UoF?

Data

The police UoF data were retrieved digitally from the SFPD for the years 2017 and 2018, with each row in the raw dataset representing one officer involved in a reported incident. However, because multiple officers and force techniques may be recorded within a single encounter, several administrative UoF records may correspond to a single incident. To analyse UoF events effectively, the data were transformed to represent individual UoF incidents. Records referring to the same encounter were merged into a single incident record, while retaining a count of the number of officers involved. When multiple force techniques were reported within the same encounter, the highest recorded force severity was retained for classification according to the SFPD (2022b).

The final dataset contained 2620 unique UoF incidents, including information on incident location, timing, and force type, and covers all recorded police–citizen encounters involving UoF, regardless of the precipitating offence type or severity. Seven UoF categories were included (see Table 1). For analytical purposes, these were grouped into two binary categories²: (1) hands-on force, involving direct physical control (e.g., physical control, strikes with object or fist, impact weapon); and (2) hands-off force, involving the deployment of a device without direct physical contact (e.g., firearm pointing, firearm discharge, and pepper spray/OC). This classification reflects the mode of force application—direct physical contact versus device-mediated force—rather than a hierarchy of severity, either between the two categories or among the techniques within each.

Table 1.

Type of use-of-force incidents reported by the SFPD.

	N	(%)	Type of UoF
Pointing of firearms	1351	51.56	Hands-off
Physical control	794	30.31	Hands-on
Strike by object or fist	313	11.95	Hands-on
Impact weapon	102	3.89	Hands-on
Pepper spray (/OC)	53	2.02	Hands-off
Firearm	7	.27	Hands-off
Total	2620 ⁵	100
Hands-on	1209	46.15
Hands-off	1411	53.85

Geocoding of all UoF incident addresses was performed using the tidygeocoder package v. 1.0.5 in R and the ArcGIS geocoding API (Cambon et al., 2022). Five incidents with missing address information were excluded from the analysis. The minimum level of geoinformation precision was the intersection level (e.g. Lincoln Ave/9th Ave); therefore, entries with only district-level geolocation were also removed. The final dataset achieved a geocoding hit rate of 99.47%.

Geocoded incidents were matched to the census block level using publicly available block geodata from https://DataSF.org (United States Census Bureau, 2024). After manually tidying the geospatial data, the final dataset included 5966 census blocks. Additional demographics from the United States Census Bureau (2020) were merged with the UoF data for descriptive analysis.

Measures and analysis

Spatial concentration and agent concentration

To evaluate the concentration of UoF on census blocks, police officers, and suspects, we calculated cumulative percentages and the Gini coefficient. As spatial research is often confronted with larger sample sizes of spatial units than event data, we calculated the Gini coefficient for all units and for all units with at least one reported incident. The Gini coefficient measures the distribution of variables across the population, with a coefficient of zero indicating a perfectly equal distribution while a value of 1 indicates complete concentration, where all the variables are concentrated in a single unit.

Temporal stability and concentration

The temporal stability of the spatial concentration of police UoF was assessed by examining the overlap of hotspots (defined as 10 or more UoF incidents) and hottest spots (15 or more UoF incidents).³ This analysis aimed to identify stable areas of high UoF concentration over time. Additionally, we analysed the temporal concentration of hotspots across different times of day.

Multi-officer and proactive vs reactive policing

To assess the outcomes of UoF incidents, we also examined the influence of multi-officer involvement – incidents involving more than one officer (Henriksen and Kruke, 2020) and proactive policing, where the officer initiated the encounter without a dispatch order versus reactive policing (for a review of proactive versus reactive policing, see Weisburd and Majimundar, 2018). Chi-square tests and odds ratios were calculated to assess the following outcomes: (1) the occurrence of officer injury, (2) the occurrence of suspect injury, and, (3) hands-off force by police officers. Notably, injuries were obtained directly from the SFPD UoF reporting system and recorded as binary indicators (injury vs no injury), regardless of injury type or severity⁴, as coded in the original dataset.

Minority concentration vs UoF concentration

Given the frequent examination of racial factors and potential biases concerning the prevalence of police stops and UoF incidents (Smith et al., 2022), we mapped population data to census blocks in San Francisco. This mapping aimed to visually explore the geographic concentration of minority groups compared to the concentration of UoF, focusing on Black and Asian residents. To identify minority neighbourhoods or census blocks, they were classified into four different categories, emphasising blocks and clusters where the majority of the population is either Black or Asian, with proportions ranging from 50% to over 75%.

Results

Spatial, agent, and temporal concentrations

Police use of force is highly concentrated at the micro-geographic level (see Table 2). A quarter of all UoF incidents occur in 1.27% of census blocks (n = 76), and half of all incidents take place in 4.41% of blocks (n = 263). Only 22.78% of census blocks reported any UoF incidents. The cumulative concentration is lower across non-zero census blocks with at least one incident (25^th percentile of UoF: 5.59%; 50^th percentile of UoF: 19.35%). The Gini coefficient (GΦ) for all blocks is .860, indicating a highly skewed distribution of UoF incidents across all blocks. GΦ for non-zero blocks is .387, suggesting a less concentrated but still unequal distribution, with more blocks involved in the incidents. These high-concentration UoF incident blocks were primarily located along a central avenue in the northeast of San Francisco (see Figure 1).

Table 2.

The concentration of use-of-force incidents across blocks.

	N	% of all blocks	% of blocks with at least one UoF incident
Total number of blocks	5966	100
Blocks with no UoF incidents	4607	77.22
Blocks with 25% of UoF incidents	76	1.27	5.59
Blocks with 50% of UoF incidents	263	4.41	19.35
Gini coefficient: .860
Gini coefficient^Φ: .387 (non-zero blocks)

Figure 1.

Use-of-force incidents per census block in San Francisco for the years 2017 and 2018.

Agent-level concentration of UoF is lower than spatial concentration in census blocks, both for officers and suspects, and whether considering all blocks or only those with recorded incidents (see Table 3). Approximately 6% of police officers (n = 74) accounted for 25% of UoF incidents, while 71% of officers (n = 870) were involved in more than one UoF incident during the study period. The GΦ for officers is .494, indicating moderate concentration among a subset of officers. Concentration among suspects is lower still: approximately 9% of suspects accounted for 25% of incidents, with a GΦ of .321, reflecting a more dispersed distribution of UoF incidents across individuals.

Table 3.

The concentration of use-of-force incidents among officers and suspects.

	Police officers		Suspects
	N	% of officers	N	% of suspects
Total number of agents	1226	100	3143	100
Agents with 25% of UoF incidents	74	6.04	272	8.65
Agents with 50% of UoF incidents	220	17.95	844	26.85
Agents with only one UoF incident	356	29.04	1665	52.98
	Gini coefficient: .494		Gini coefficient: .321

Temporally, the probability of UoF occurrence follows a bimodal distribution, peaking at 4 pm and 11 pm. However, these peaks vary by spatial concentration level (see Figure 2). In hotspots (10 UoF incidents), the highest probabilities occur at 1 pm and 3 pm; in the hottest spots (15 or more incidents), they shift to 4 pm and 8 pm. Across all categories, UoF probability is lowest between 5 am and 9 am.

Figure 2.

Probability of use-of-force occurrence by hour across all places, hotspots, and hottest spots.

Hotspot stability is quite low. Three census blocks classified as hotspots in 2017 remained hot in 2018, representing a spatiotemporal overlap of 15% (see Table 4). No overlap for the hottest spots in 2017 and 2018 is reported (Dau et al., 2022).

Table 4.

Overlap of hotspots and hottest spots in 2017 and 2018.

	n	%
Hotspots₂₀₁₇ ∩ Hotspots₂₀₁₈	3	15
Hotspots₂₀₁₇ ∩ Hotspots_Total	12	60
Hotspots₂₀₁₈ ∩ Hotspots_Total	10	50
Hottest spots₂₀₁₇ ∩ Hottest spots₂₀₁₈	0	0
Hottest spots₂₀₁₇ ∩ Hottest spots_Total	3	15
Hottest spots₂₀₁₈ ∩ Hottest spots_Total	2	10

The geography of UoF

We mapped the census block, which was identified as the UoF hotspots and hottest spots for 2017 and 2018 in San Francisco. All spots were clustered around or in close proximity to Market Street, which dissects the northeastern part of San Francisco, with some hotspots being dispersed south of the avenue. Twelve hotspots were identified in 2017 and 13 in 2018 (see Figure 3). Five unique hottest spots were identified in 2017 and four in 2018. One hotspot from 2017 turned into the hottest spot in 2018. The hottest spots during both years show a more robust spatial clustering along the central avenue than the hotspots.

Figure 3.

Hotspots and hottest spots in San Francisco for the years 2017 and 2018.

In addition to the occurrence of UoF, we mapped the outcomes of UoF in terms of injuries sustained by police officers and suspects during UoF incidents. During 2017 and 2018, police officers were injured in 9.6% of UoF incidents (n = 253), while suspects were injured in 32.7% of incidents (n = 858). Census blocks with a high number of UoF incidents also exhibited a high number of officer injuries, particularly on the east side of Market Street—a major avenue in the northeast of San Francisco. Clusters of officer injuries were evident across the city’s downtown districts (see Figure 4). Meanwhile, suspect injuries demonstrated a more pronounced clustering effect along the avenue and in the Mission District, with stronger clustering than officer injuries (see Figure 5). The block with the highest number of reported suspect injuries recorded 10 incidents during the study period. This difference in clustering is partly due to the larger sample size of suspect injuries, which is more than three times that of police officer injuries.

Figure 4.

Officer injuries per census block in San Francisco for the years 2017 and 2018.

Figure 5.

Suspect injuries per census block in San Francisco for the years 2017 and 2018.

Effects of multi-officer and proactive policing on UoF

We used chi-square tests to identify differences in the occurrence of officer and suspect injuries and in the use of hands-off force between encounters involving a single officer and those involving multiple officers. Single-officer encounters where force was applied were significantly associated with fewer officer injuries. While 12.96% of multiple-officer encounters in which force was used led to an officer injury, 6.76% of single-officer encounters resulted in an officer injury (see Table 5). Suspect injury and hands-off force were also reported to occur significantly less often during single-officer encounters involving police UoF.

Table 5.

Chi-square test for comparing single and multiple-officer involvement in use-of-force Incidents.

	Single officer		Multiple officers		x ²	p
	N	(%)	N	(%)	x ²	p
Officer injury**					28.09	.000
Yes	95	6.76	158	12.96
No	1310	93.24	1061	87.04
Suspect injury**					10.01	.002
Yes	421	29.96	437	35.85
No	984	70.04	782	64.15
Type of force*					4.84	.028
Hands-on	678	48.26	535	43.89
Hands-off	727	51.74	684	56.11

*p < .05, **p < .01.

Odds ratios were used to calculate the proportional risk of injury and hands-off force occurrence. All differences between single-officer and multi-officer encounters are reported to be statistically significant (p < .05). As shown in Table 6, single-officer encounters have a 51% lower risk of resulting in officer injuries, a 23% lower risk of suspect injuries, and a 15% lower risk of hands-off UoF.

Table 6.

Odds ratio estimates on the difference between single and multi-officer involvement on the risk of officer injury, suspect injury, and the type of force used.

	Single officer
	Estimate	Lower CI	upper CI	p
Officer injury**	.487	.372	.635	.000
Suspect injury**	.766	.650	.902	.001
Type of force*	.839	.719	.978	.025

*p < .05, **p < .01.

The chi-square test of proactive and reactive encounters revealed no statistically significant differences across any of the three variables (see Table 7). The most considerable difference is found for the effect on suspect injury. Approximately three in 10 proactive encounters (30.72%) that involved police UoF resulted in suspect injury. The prevalence is slightly higher for reactive encounters (34.16%), though this difference is only marginally statistically significant (p = .069).

Table 7.

Chi-square test for comparing proactive and reactive policing in use-of-force incidents.

	Proactive policing		Reactive policing		x ²	p
	N	(%)	N	(%)	x ²	p
Officer injury					.007	.931
Yes	108	9.5	143	9.68
No	1028	90.5	1343	90.32
Suspect injury					3.11	.069
Yes	349	30.72	507	34.16
No	787	69.28	979	65.84
Type of force					.360	.548
Hands-on	533	46.92	679	45.66
Hands-off	603	53.08	808	54.34

*p < .05, **p < .01.

Moreover, odds ratios were used to assess the risks associated with proactive and reactive policing, as shown in Table 8. Overall, proactive encounters appeared to be associated with a slightly lower likelihood of officer or suspect injury and a lower likelihood of hands-off force compared to reactive encounters. Specifically, the risk of suspect injury was 14.5% lower (OR = .855; 95% CI = .724–1.01; p = .063).

Table 8.

Odds ratio estimates of the difference between proactive and reactive policing on the risk of officer injury, suspect injury, and the type of force used.

	Proactive policing
	Estimate	Lower CI	Upper CI	p
Officer injury	.881	.753	1.27	.367
Suspect injury	.855	.724	1.01	.063
Type of force	.951	.814	1.11	.522

*p < .05, **p < .01.

Geographic interdependence of UoF and the minority population

We also examined the geographic concentration of minority groups in relation to UoF incidents. As shown in Figure 6, census blocks with a majority Black population are sparse and dispersed, with only 5–10 individual blocks containing more than 50% Black residents, located primarily in the northeastern and southeastern parts of the city. The only visible cluster appears around the Hunter Point district. Comparing Figures 1 and 3 with Figure 6 reveals that census blocks with high proportions of Black residents (>50%) do not overlap with those identified as UoF hotspots or hottest spots (0% overlap).

Figure 6.

Proportion of black population per census block in San Francisco for the years 2017 and 2018.

As shown in Figure 7, a substantial number of census blocks (approximately 100–150) have more than 50% Asian residents, with some exceeding 75%. The strongest cluster is located, as expected, in Chinatown at the northeastern edge of the city, with another significant cluster in the southeast near the Silver Terrace area. As with the Black population, comparing Figures 1 and 3 with Figure 7 reveals that census blocks with a majority Asian population (>50%) do not overlap with those identified as UoF hotspots or hottest spots (0% overlap).

Figure 7.

Proportion of Asian population per census block in San Francisco for the years 2017 and 2018.

Discussion

This study provides the first combined geographic and temporal analysis of police use of force at the micro-level. It examines the characteristics of police encounters that result in officer or suspect injuries or hands-off force. Compared to other research (Braga and Bond, 2008; Braga et al., 2011; Taylor et al., 2011, 2024), the concentration in this study was not limited to spatial dimensions but also included temporal and agent-based dimensions. Our findings reveal that police UoF is highly concentrated at the micro-geographic level. While the concentration among officers and suspects is less pronounced than the spatial concentration within census blocks, they still demonstrated a meaningful level of concentration. Additionally, the frequency of use-of-force incidents varies throughout the day, with a significant temporal concentration during certain hours, while the spatial hotspots show instability, with little to no overlap between 2017 and 2018. We also observed that both police officers and suspects sustained injuries during these incidents, but suspect injuries were more numerous and clustered.

Following Sorg et al. (2021), UoF incidents exhibit a similar spatial concentration as crime. Similar levels of spatial concentration might be partly a natural consequence of spatial overlap between crime and police UoF. Places with high levels of crime due to certain crime attractors (Kinney et al., 2008; Weisburd et al., 2016: p. 48) tend to attract more potential offenders, which, in turn, may lead to increased police encounters within these places, which might result in a higher prevalence of police UoF.

Beyond the prevailing focus on offenders and, lately, increasingly on places, this study identified a notable concentration of UoF incidents among officers. This pattern reflects a broader principle of distributional inequality observed in crime analysis, where a minority of units (officers, places, or offenders) generate the majority of outcomes. The spatial concentration of UoF may be influenced by the deployment of high-use officers in specific areas, highlighting the need for future research to examine the extent to which UoF patterns are shaped by officer deployment, individual officer behaviour, or a combination of both.

Furthermore, we found that encounters with individual officers were associated with lower risks of injury and hands-off force than incidents involving multiple officers. This association may, in part, reflect that multi-officer encounters are more likely to occur in complex or higher-risk situations, where the probability of injury is already elevated. At the same time, prior research has suggested that the likelihood of UoF increases with the presence of more officers (Bolger, 2015; Henriksen and Kruke, 2020). As Paoline (2003) argued, the perception of the officers’ work environment and tasks as dangerous and threatening may unite officers, reinforcing a cultural mandate of loyalty that drives them to support one another both physically and emotionally. This culture, characterised by widely shared attitudes, values, and norms, helps manage the strains of police work and punitive management practices. However, it often involves distrust and suspicion of citizens, a focus on maintaining authority, a ‘us versus them’ mentality, and intense loyalty to peers. Such an environment, which fosters a collective response, may escalate the use of force when additional officers are involved.

Similarly, identifying hotspots of police use of force and concentrating resources, such as increased supervision and oversight, in these areas could help manage the risk of UoF and potentially reduce the frequency of UoF. By targeting efforts at locations where police UoF is most prevalent, it is possible to manage and reduce UoF in these hotspots, thereby helping control citywide UoF (Weisburd et al., 2017). This targeted strategy aligns with the broader ‘hot spots policing’ approach, which emphasises concentrating resources in high-activity areas to achieve effective results (Braga et al., 2014, 2019a, 2019b; Braga and Weisburd, 2022; Weisburd et al., 2024; Weisburd and Telep, 2014).

Geographic mapping precision and overlap

Compared to crime, the spatial geographies of the police UoF remain very much underexplored. Interestingly, fundamental mapping of hotspots is not a common practice in research on policing and its effects. Understanding the spatial concentration of any phenomenon represents the first step in investigating the factors associated with its (re)occurrence. This study has shown that hotspots and hottest spots of UoF are concentrated in a few census blocks. Mapping the specific locations where police use of force occurs can help guide specially trained personnel to these locations to better respond to place-based risk and prevent UoF.

Moreover, focusing on the highest spatial resolution should become the default in place-based research. This also includes basic maps of the phenomenon of interest. There are two reasons for this. First, collecting data at the most precise level (i.e., street segments or blocks) allows aggregation to any lower spatial resolution (i.e., neighbourhood or city level). Second, determining the precise locations where police UoF occurs offers practical insights to police chiefs and strategists to better guide patrol officers and analyse place-based differences.

Limitations

This study examines reported UoF incidents in the SFPD during 2017–2018 and therefore reflects a single organisational and temporal context. As such, the findings have limited generalisability, but the exploratory design offers a framework that may support future comparative and cumulative research. Second, the analysis relies on census-based residential population data, but residential location may be a poor proxy for the places in which offenders conduct their routine activities (see also Basta et al., 2010; Colburn et al., 2020; Jones and Pebley, 2014). Mobility-based measures could better capture the geographic overlap between police activity and the presence of different racial groups. Relatedly, while the study benchmarks UoF outcomes directly against census populations and examines encounter-level reactivity by comparing proactive and reactive encounters, it does not incorporate earlier stages in the police–citizen process (e.g., who is reported to the police, calls for service, stops, arrests), where disparities may already emerge and shape exposure to encounters in which UoF may occur. Future research could address these limitations by incorporating mobility-based population measures and earlier stages of police–citizen interaction, and by applying bivariate spatial visualisations to further examine potential spatial co-occurrence.

Third, while “non-conforming” handcuffs are generally considered a step on the continuum of force (Henstock and Ariel, 2017), our study did not include handcuffing in the measurement of UoF. This decision was based on the data provided by SFPD, which did not categorise handcuffing as part of UoF in their reporting structure (For a discussion on the inconsistency in how handcuffing is treated across different departments, see Henstock and Ariel, 2017). Although our approach aligns with previous studies that exclude handcuffs (e.g., Assaraf et al., 2024) and offers a consistent, comparable measure, excluding handcuffs may lead to an underestimation of UoF incidents.

Finally, injury outcomes were examined at the incident level, reporting overall proportions and spatial concentrations of injury rather than block-specific injury probabilities. Future research should model block-level injury rates to identify places where the likelihood of injury is disproportionately high relative to incident volume. In addition, future studies should account for the type of offence that led to the encounter, as different encounter contexts may differentially shape both the likelihood of UoF and the probability of injury.

Conclusions

This study demonstrates that police use of force is highly concentrated in specific places, times, and among certain officers. The place-based nature of UoF offers new perspectives and actionable insights for researchers and practitioners seeking to evaluate and design hotspot policing programmes. By combining place-based frameworks (focusing on where UoF occurs) with agent-based approaches (focusing on who is involved), police departments can identify high-risk locations and officers, enabling targeted interventions in areas where the likelihood of force is elevated. Understanding the temporal and spatial distribution of UoF incidents can also inform decisions regarding officer deployment and the allocation of training resources. To operationalise these insights sustainably, data collection and analysis should extend beyond spatial factors to incorporate information on the when, where, and who of policing activities. This triangulation enables a more comprehensive understanding of policing outcomes in specific local contexts. Finally, this study provides the first detailed geographic mapping of UoF incidents and recommends integrating richer data sources—such as mobility data—to enhance future analyses.

Footnotes

Acknowledgements

We would like to express our sincere gratitude to Principal Analyst Dae Son and Deputy Chief Peter D. Walsh from the San Francisco Police Department for providing the data and facilitating the data-sharing process that made this research possible.

ORCID iDs

Noy Assaraf

Philipp M. Dau

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Notes

References

Adams

Alpert

(2023) Use of Force in Policing. Oxford research encyclopedia of criminology and criminal justice. https://doi.org/10.1093/acrefore/9780190264079.013.845

Alpert

Dunham

(1998) Force Factor: Measuring Police Use of Force Relative to Suspect Resistance. Washington, DC: Police Executive Research Forum. https://www.ojp.gov/library/publications/force-factor-measuring-police-use-force-relative-suspect-resistance

Alpert

Smith

(1994) How reasonable is the reasonable man: police and excessive force. Journal of Criminal Law and Criminology 85: 481–501. https://doi.org/10.2307/1144107

Ariel

Partridge

(2017) Predictable policing: measuring the crime control benefits of hotspots policing at bus stops. Journal of Quantitative Criminology 33(4): 809–833. https://doi.org/10.1007/s10940-016-9312-y

Ariel

Farrar

Sutherland

(2015) The effect of police body-worn cameras on use of force and citizens’ complaints against the police: a randomized controlled trial. Journal of Quantitative Criminology 31(3): 509–535. https://doi.org/10.1007/s10940-014-9236-3

Ariel

Sutherland

Henstock

, et al. (2016a) Report: increases in police use of force in the presence of body-worn cameras are driven by officer discretion: a protocol-based subgroup analysis of ten randomized experiments. Journal of Experimental Criminology 12(3): 453–463. https://doi.org/10.1007/s11292-016-9261-3

Ariel

Weinborn

Sherman

(2016b) ‘Soft’ policing at hot spots—do police community support officers work? A randomized controlled trial. Journal of Experimental Criminology 12(3): 277–317. https://doi.org/10.1007/s11292-016-9260-4

Ariel

Langton

Peters

, et al. (2024) Private security for curbing unwanted sexual behaviours in train stations: a place-based randomised controlled trial. Journal of Experimental Criminology 21: 1325–1353. https://doi.org/10.1007/s11292-024-09632-x

Assaraf

Mouro

Papy

, et al. (2024) Behind the yellow sticker: paradoxical effects of a visual warning of body-worn cameras on the use of police force. Journal of Experimental Criminology 21: 1215–1237. https://doi.org/10.1007/s11292-024-09627-8

10.

Assaraf

Mouro

Papy

, et al. (2026) Research note: Body-worn camera footage analysis of use-of-force encounters—A closer look at the paradoxical effect of visual warnings. Journal of Experimental Criminology. Available at: https://doi.org/10.1007/s11292-026-09748-2

11.

Basta

Richmond

Wiebe

(2010) Neighborhoods, daily activities, and measuring health risks experienced in urban environments. Social Science & Medicine 71(11): 1943–1950. https://doi.org/10.1016/j.socscimed.2010.09.008

12.

Bolger

(2015) Just following orders: a meta-analysis of the correlates of American police officer use of force decisions. American Journal of Criminal Justice 40(3): 466–492. https://doi.org/10.1007/s12103-014-9278-y

13.

Braga

Bond

(2008) Policing crime and disorder hot spots: a randomized controlled trial. Criminology 46(3): 577–607. https://doi.org/10.1111/j.1745-9125.2008.00124.x

14.

Braga

Weisburd

(2022) Does hot spots policing have meaningful impacts on crime? Findings from an alternative approach to estimating effect sizes from place-based program evaluations. Journal of Quantitative Criminology 38(1): 1–22. https://doi.org/10.1007/s10940-020-09481-7

15.

Braga

Hureau

Papachristos

(2011) An ex post facto evaluation framework for place-based police interventions. Evaluation Review 35(6): 592–626. https://doi.org/10.1177/0193841X11433827

16.

Braga

Papachristos

Hureau

(2014) The effects of hot spots policing on crime: an updated systematic review and meta-analysis. Justice Quarterly 31(4): 633–663. https://doi.org/10.1080/07418825.2012.673632

17.

Braga

Turchan

Papachristos

, et al. (2019a) Hot spots policing of small geographic areas effects on crime. Campbell Systematic Reviews 15(3): e1046. https://doi.org/10.1002/cl2.1046

18.

Braga

Turchan

Papachristos

, et al. (2019b) Hot spots policing and crime reduction: an update of an ongoing systematic review and meta-analysis. Journal of Experimental Criminology 15(3): 289–311. https://doi.org/10.1007/s11292-019-09372-3

19.

Brandl

Stroshine

(2013) The role of officer attributes, job characteristics, and arrest activity in explaining police use of force. Criminal Justice Policy Review 24(5): 551–572. https://doi.org/10.1177/0887403412452424

20.

Cambon

Hernangómez

Belanger

, et al. (2022) Tidygeocoder: geocoding made easy. R package version 1.0.5. https://doi.org/10.32614/CRAN.package.tidygeocoder

21.

Cojean

Combalbert

Taillandier-Schmitt

(2020) Psychological and sociological factors influencing police officers’ decisions to use force: a systematic literature review. International Journal of Law and Psychiatry 70: 101569. https://doi.org/10.1016/j.ijlp.2020.101569

22.

Colburn

Pratt

Mueller

, et al. (2020) How adolescents define their home neighborhoods conceptually and spatially. Journal of Community Psychology 48(3): 709–725. https://doi.org/10.1002/jcop.22290

23.

Cowell

Corsi

Johnson

, et al. (2021) The factors that motivate law enforcement’s use of force: a systematic review. American Journal of Community Psychology 67(1): 142–151. https://doi.org/10.1002/ajcp.12460

24.

Dau

Dewinter

Witlox

, et al. (2022) How concentrated are police on crime? A spatiotemporal analysis of the concentration of police presence and crime. Cambridge Journal of Evidence-Based Policing 6(3): 109–133. https://doi.org/10.1007/s41887-022-00079-6

25.

Durán

Loza

(2017) Exploring the two trigger fingers thesis: racial and ethnic differences in officer involved shootings. Contemporary Justice Review 20(1): 71–94. https://doi.org/10.1080/10282580.2016.1262771

26.

FRED (2023) Per capita personal income in San Francisco county/city, CA. Federal Reserve Economic Data. https://fred.stlouisfed.org/series/PCPI06075

27.

Fryer

(2019) An empirical analysis of racial differences in police use of force. Journal of Political Economy 127(3): 1210–1261. https://doi.org/10.1086/701423

28.

Gill

Wooditch

Weisburd

(2017) Testing the ‘law of crime concentration at place’ in a suburban setting: implications for research and practice. Journal of Quantitative Criminology 33(3): 519–545. https://doi.org/10.1007/s10940-016-9304-y

29.

Gorr

Lee

(2012) Longitudinal study of Crime Hot Spots: Dynamics and Impact on Part 1 Violent Crime. University of South Florida. https://doi.org/10.13140/RG.2.1.2677.1365/1

30.

Groff

Weisburd

Yang

S-M

(2010) Is it important to examine crime trends at a local ‘micro’ level? a longitudinal analysis of street to street variability in crime trajectories. Journal of Quantitative Criminology 26(1): 7–32. https://doi.org/10.1007/s10940-009-9081-y

31.

Henriksen

Kruke

(2020) The force continuum: prevalence and characteristics of police use of coercive force. Nordic Journal of Studies in Policing 7(1): 5–22. https://doi.org/10.18261/issn.2703-7045-2020-01-02

32.

Henstock

Ariel

(2017) Testing the effects of police body-worn cameras on use of force during arrests: a randomised controlled trial in a large British police force. European Journal of Criminology 14(6): 720–750. https://doi.org/10.1177/1477370816686120

33.

Hickman

Atherley

Lowery

, et al. (2015) Reliability of the force factor method in police use-of-force research. Police Quarterly 18(4): 368–396. https://doi.org/10.1177/1098611115586175

34.

Hine

Porter

Westera

, et al. (2018) Too much or too little? Individual and situational predictors of police force relative to suspect resistance. Policing and Society 28(5): 587–604. https://doi.org/10.1080/10439463.2016.1232257

35.

Holmes

Painter II

Smith

(2019) Race, place, and police-caused homicide in U.S. municipalities. Justice Quarterly 36(5): 751–786. https://doi.org/10.1080/07418825.2018.1427782

36.

International Association of the Chiefs of Police (2001) Police use of force in America, 2001. https://www.theiacp.org/sites/default/files/2018-08/2001useofforce.pdf

37.

Jain

Sinclair

Papachristos

(2022) Identifying misconduct-committing officer crews in the Chicago police department. PLoS One 17(5): e0267217. https://doi.org/10.1371/journal.pone.0267217

38.

Johnson

Lab

Bowers

(2008) Stable and fluid hotspots of crime: differentiation and identification. Built Environment 34(1): 32–45. https://doi.org/10.2148/benv.34.1.32

39.

Jones

Pebley

(2014) Redefining neighborhoods using common destinations: social characteristics of activity spaces and home census tracts compared. Demography 51(3): 727–752. https://doi.org/10.1007/s13524-014-0283-z

40.

Kim

(2018) Examining the relationship between the structural characteristics of place and crime by imputing census block data in street segments: is the pain worth the gain? Journal of Quantitative Criminology 34: 67–110. https://doi.org/10.1007/s10940-016-9323-8

41.

Kinney

Brantingham

Wuschke

, et al. (2008) Crime attractors, generators and detractors: land use and urban crime opportunities. Built Environment 34(1): 62–74. https://doi.org/10.2148/benv.34.1.62

42.

Klahm

Tillyer

(2010) Understanding police use of force: a review of the evidence. Southwest Journal of Criminal Justice 7: 214–239.

43.

Kumar

Ariel

Hodgkinson

, et al. (2024) ‘Domestic abuse hot spots’: a longitudinal, place-based analysis of 13 years of initial reports to the police. Policing: Journal of Policy Practice 18: paae056. https://doi.org/10.1093/police/paae056

44.

Lautenschlager

Omori

(2019) Racial threat, social (dis)organization, and the ecology of police: towards a macro-level understanding of police use-of-force in communities of color. Justice Quarterly 36(6): 1050–1071. https://doi.org/10.1080/07418825.2018.1480792

45.

Lersch

(2022) An exploratory spatiotemporal analysis of domestic disturbance calls for service in 2020: the case of Tampa, Florida, USA. International Journal of Comparative and Applied Criminal Justice 48: 1–16. https://doi.org/10.1080/01924036.2022.2144396

46.

Lersch

Bazley

Mieczkowski

, et al. (2008) Police use of force and neighbourhood characteristics: an examination of structural disadvantage, crime, and resistance. Policing and Society 18(3): 282–300. https://doi.org/10.1080/10439460802091690

47.

Longridge

Chapman

Bennell

, et al. (2023) Behaviour sequence analysis of police body-worn camera footage. Journal of Police and Criminal Psychology 38(2): 255–262. https://doi.org/10.1007/s11896-020-09393-z

48.

MacDonald

Fagan

Geller

(2016) The effects of local police surges on crime and arrests in New York city. PLoS One 11(6): e0157223. https://doi.org/10.1371/journal.pone.0157223

49.

Ouellet

Hashimi

Gravel

, et al. (2019) Network exposure and excessive use of force. Criminology & Public Policy 18(3): 675–704. https://doi.org/10.1111/1745-9133.12459

50.

Paoline

(2003) Taking stock: toward a richer understanding of police culture. Journal of Criminal Justice 31(3): 199–214. https://doi.org/10.1016/S0047-2352(03)00002-3

51.

Piza

O’Hara

(2014) Saturation foot-patrol in a high-violence area: a quasi-experimental evaluation. Justice Quarterly 31(4): 693–718. https://doi.org/10.1080/07418825.2012.668923

52.

Piza

Connealy

Sytsma

, et al. (2023) Situational factors and police use of force across micro-time intervals: a video systematPIZAic social observation and panel regression analysis. Criminology 61(1): 74–102. https://doi.org/10.1111/1745-9125.12323

53.

Ratcliffe

(2004) The hotspot matrix: a framework for the spatio‐temporal targeting of crime reduction. Police Practice and Research 5(1): 5–23. https://doi.org/10.1080/1561426042000191305

54.

Sampson

(2013) The place of context: a theory and strategy for criminology’s hard problems. Criminology 51(1): 1–31. https://doi.org/10.1111/1745-9125.12002

55.

Santos

(2015) An ex post facto evaluation of tactical police response in residential theft from vehicle micro-time hot spots. Journal of Quantitative Criminology 31(4): 679–698. https://doi.org/10.1007/s10940-015-9248-7

56.

Santos

(2016) Offender-focused police intervention in residential burglary and theft from vehicle hot spots: a partially blocked randomized control trial. Journal of Experimental Criminology 12(3): 373–402. https://doi.org/10.1007/s11292-016-9268-9

57.

Schnell

McManus

(2022) The influence of temporal specification on the identification of crime hot spots for program evaluations: a test of longitudinal stability in crime patterns. Journal of Quantitative Criminology 38(1): 51–74. https://doi.org/10.1007/s10940-020-09483-5

58.

Sernoffsky

(2018) SF Police Commission Approves TASER Policy for Officers. Police1. https://www.police1.com/police-products/less-lethal/taser/articles/sf-police-commission-approves-taser-policy-for-officers-68CylCzZeF5PxQyG/

59.

SF.GOV (2018) Use of Tasers by San Francisco Police Officers. https://sfelections.sfgov.org/sites/default/files/Documents/candidates/TitleAndSummary_UseOfTasersBySanFranciscoPoliceOfficers.pdf

60.

SFPD (2022a) Staffing analysis of the San Francisco police department. City & County of San Francisco Police Department. https://www.sanfranciscopolice.org/sites/default/files/2022-03/SFPD2021StaffingAnalysisReport20220307.pdf

61.

SFPD (2022b) Department general order 5.01: use of force policy and proper control of a person. San Francisco Police Department. https://www.sf.gov/sites/default/files/2022-09/PoliceCommission090722-DGO/Use/of/Force/Policy/Proper?Control/of?a/Person/Clean/Copy.pdf

62.

SFPD (2024b) Station finder. San Francisco Police Department. https://www.sanfranciscopolice.org/your-sfpd/sfpd-stations/station-finder

63.

Sherman

(1980) Causes of police behavior: the current state of quantitative research. Journal of Research in Crime and Delinquency 17(1): 69–100. https://doi.org/10.1177/002242788001700106

64.

Sherman

Weisburd

(1995) General deterrent effects of police patrol in crime ‘hot spots’: a randomized, controlled trial. Justice Quarterly 12(4): 625–648. https://doi.org/10.1080/07418829500096221

65.

Sherman

Gartin

Buerger

(1989) Hot spots of predatory crime: routine activities and the criminology of place. Criminology 27(1): 27–56. https://doi.org/10.1111/j.1745-9125.1989.tb00862.x

66.

Short

Brantingham

Bertozzi

, et al. (2010) Dissipation and displacement of hotspots in reaction-diffusion models of crime. Proceedings of the National Academy of Sciences 107(9): 3961–3965. https://doi.org/10.1073/pnas.0910921107

67.

Smith

Ferguson

San Miguel

, et al. (2022) Social, health and ethnicity correlates of complaints of excessive police force. Criminal Behaviour and Mental Health 32(4): 284–294. https://doi.org/10.1002/cbm.2255

68.

Smith

Tillyer

Tregle

(2024) Hot spots policing as part of a city-wide violent crime reduction strategy: initial evidence from Dallas. Journal of Criminal Justice 90: 102091. https://doi.org/10.1016/j.jcrimjus.2023.102091

69.

Sorg

Houser

Lewandowski

, et al. (2021) The geographic concentration of use of force incidents at subway stations mimics the law of crime concentration bandwidths: a research note. Police Journal 94(3): 316–332. https://doi.org/10.1177/0032258X20928096

70.

Taylor

Koper

Woods

(2011) A randomized controlled trial of different policing strategies at hot spots of violent crime. Journal of Experimental Criminology 7(2): 149–181. https://doi.org/10.1007/s11292-010-9120-6

71.

Taylor

Liu

Maitra

, et al. (2024) The effects of community-infused problem-oriented policing in crime hot spots based on police data: a randomized controlled trial. Journal of Experimental Criminology 20(2): 317–345. https://doi.org/10.1007/s11292-022-09541-x

72.

Terrill

Zimmerman

Somers

(2023) Applying video-based systematic social observation to police use of force encounters: an assessment of de-escalation and escalation within the context of proportionality and incrementalism. Justice Quarterly 40: 1–30. https://doi.org/10.1080/07418825.2023.2222819

73.

Tillyer

(2022) Unpacking sequential actions within use of force incidents. Police Quarterly 25(2): 178–200. https://doi.org/10.1177/10986111211049549

74.

Townsley

(2008) Visualising space-time patterns in crime: the hotspot plot. Crime Patterns and Analysis 1(1): 61–74.

75.

Tuch

Weitzer

(1997) Trends: racial differences in attitudes toward the police. Public Opinion Quarterly 61(4): 642–663. https://doi.org/10.1086/297822

76.

United States Census Bureau (2020) Total Population. 2020: DEC Demographic and Housing Characteristics. https://data.census.gov/cedsci/table?q=Population&g=0400000US06_0500000US06075,06075%241000000&y=2020

77.

United States Census Bureau (2023) Quickfacts, San Francisco City, California; San Francisco County, California. An official website of the United States government. https://www.census.gov/quickfacts/fact/table/sanfranciscocitycalifornia,sanfranciscocountycalifornia/PST045221

78.

United States Census Bureau (2024) Census 2020: block groups for San Francisco, geographic locations and boundaries. Data. https://data.sfgov.org/Geographic-Locations-and-Boundaries/Census-2020-Block-Groups-for-San-Francisco/24e8-pd2q/about_data

79.

Weinborn

Ariel

Sherman

, et al. (2017) Hotspots vs. harmspots: shifting the focus from counts to harm in the criminology of place. Applied Geography 86: 226–244. https://doi.org/10.1016/j.apgeog.2017.06.009

80.

Weisburd

(2015) The law of crime concentration and the criminology of place. Criminology 53(2): 133–157. https://doi.org/10.1111/1745-9125.12070

81.

Weisburd

(2018) Hot spots of crime and place-based prevention. Criminology & Public Policy 17(1): 5–25. https://doi.org/10.1111/1745-9133.12350

82.

Weisburd

Majimundar

(eds) (2018) Proactive Policing: Effects on Crime and Communities. The National Academies Press. https://doi.org/10.17226/24928

83.

Weisburd

Telep

(2014) Hot spots policing: what we know and what we need to know. Journal of Contemporary Criminal Justice 30(2): 200–220. https://doi.org/10.1177/1043986214525083

84.

Weisburd

Bushway

Lum

, et al. (2004) Trajectories of crime at places: a longitudinal study of street segments in the city of Seattle. Criminology 42(2): 283–322. https://doi.org/10.1111/j.1745-9125.2004.tb00521.x

85.

Weisburd

Bernasco

Bruinsma

GJN

(2009a) Putting Crime in Its Place. Springer. https://doi.org/10.1007/978-0-387-09688-9

86.

Weisburd

Morris

Groff

(2009b) Hot spots of juvenile crime: a longitudinal study of arrest incidents at street segments in Seattle, Washington. Journal of Quantitative Criminology 25(4): 443–467. https://doi.org/10.1007/s10940-009-9075-9

87.

Weisburd

Groff

Yang

S-M

(2012) The Criminology of Place: Street Segments and Our Understanding of the Crime Problem. Oxford: Oxford University Press.

88.

Weisburd

Eck

Braga

, et al. (2016) Place matters: criminology for the twenty-first century. https://doi.org/10.1017/CBO9781139342087

89.

Weisburd

Braga

Groff

, et al. (2017) Can hot spots policing reduce crime in urban areas? An agent-based simulation. Criminology 55(1): 137–173. https://doi.org/10.1111/1745-9125.12131

90.

Weisburd

Zastrow

Kuen

, et al. (2024) Crime concentrations at micro places: a review of the evidence. Aggression and Violent Behavior 78: 101979. https://doi.org/10.1016/j.avb.2024.101979

91.

Wortley

Mazerolle

(2008) Environmental criminology and crime analysis: situating the theory, analytic approach and application Environmental Criminology and Crime Analysis. Willan.