Geospatial mapping of crimes against women in West Bengal: Patterns, clusters, and policy insights

Introduction

Crimes against women (CAW) refer to acts that physically, sexually, or psychologically harm women based on their gender. Internationally, the 1993 United Nations (UN) Declaration on the Elimination of Violence Against Women (VAW) broadly defined "VAW as gender-based acts causing or likely to cause physical, sexual, or psychological harm, including threats, coercion, or arbitrary deprivation of liberty, whether in public or private life" (United Nations General Assembly, 1993) [1] . Similarly, the Council of Europe (Istanbul Convention) defines VAW as a human rights violation and a form of discrimination against women. It refers to all acts of violence that result in or are likely to result in physical, sexual, psychological, or economic harm or suffering to women, including threats, coercion, or arbitrary deprivation of liberty, whether occurring in public or private spaces (Council of Europe, 2011) [2] . While international conventions define VAW as a human rights concern, it is crucial to understand how these global principles apply at the national level. In this study, VAW and CAW are treated as synonymous, referring to legally recognized offences under the Indian Penal Code (IPC) and Special and Local Laws (SLL). In India, constitutional provisions and societal dynamics uniquely shape the contours of CAW. The Indian Constitution is the cornerstone of women's rights in India. Although the Constitution lacks a specific definition for CAW, however, Article 14 ensures equality before the law, Article 15 prohibits discrimination while allowing special provisions for women, Article 21 guarantees the right to life with dignity, and Article 51A(e) obligates citizens to reject derogatory practices against women. These constitutional provisions highlight the state's duty to legislate against crimes and prevent them. Several theoretical perspectives frame CAW as a complex issue. For instance, economic models suggest that women's financial independence and educational attainment can mitigate violence (Hampton and Gelles, 1994; Roy, 2003) [3,4] . In contrast, sociological theories propose a more intricate relationship, wherein individuals or societal structures sometimes employ violence to control women's access to resources (González-Brenes, 2004; Rowe et al. 2006) [5,6] , because it's deeply rooted in societal norms and power imbalances that are products of the social construction of crimes and deviance (Madhu, 2023) [7] . According to feminist criminologists, CAW are a systemic manifestation of patriarchal power dynamics aimed at control (Lenton, 1995; Chesney-Lind, 2006) [8,9] . India's entrenched patriarchal social structures exacerbate CAW, creating environments in which these crimes go unchallenged (Niaz, 2003) [10] . Furthermore, liberal feminists emphasize how gender inequality and biases within the criminal justice system contribute to the underreporting and inadequate addressing of offences (Malhotra et al., 1995; Madhu, 2023) [11] . Beyond these theoretical lenses, the Indian context presents a complex interplay of these perspectives within its policy and legal frameworks on CAW. Economic models support policies that promote women's education and empowerment, such as vocational training and microfinance (Paul, 2025) [12] . However, the persistence of CAW suggests that economic uplift alone cannot dismantle gender-based violence (Lolayekar et al., 2020) [13] . Sociological and feminist theories, which highlight power imbalances and societal norms, inform laws such as the Protection of Women from Domestic Violence Act (PWDVA) of 2005. This act recognizes diverse forms of domestic abuse beyond physical violence and offers civil remedies and protection orders for women. However, challenges in implementation persist, including underreporting and societal acceptance of certain forms of violence. These issues underscore the ongoing struggle against deeply rooted patriarchal social structures, as theorized by Niaz (2003) [10] . In addition to this, liberal feminist criticisms of the structural systemic biases of India's criminal justice framework remain at the centre of ongoing debates on CAW. Despite initiatives to enhance police sensitivity, judicial accountability, and legal aid accessibility, the underreporting of complaints and limited redress reflect entrenched institutional shortcomings. Legislative interventions, notably the Criminal Law (Amendment) Act of 2013, introduced progressive reforms to address women's safety issues. However, low rates of reporting, prosecution, and conviction underscore continuing structural impediments. Moving from national frameworks to specific regional manifestations, various spatial analyses have revealed significant intra-regional variations in CAW, particularly in West Bengal. For instance, studies on districts like South 24 Parganas attribute the high incidence of crimes to factors such as illiteracy and entrenched social customs (Karmakar, 2013) [14] . The range of crimes encompasses widespread physical and psychological violence, with post-marital cruelty and dowry-related abuse being the most prevalent forms of victimization in the district (Prasad, 2018; Mukherjee, 2020) [15,16] . However, these studies often employ singular spatial methods or descriptive approaches, which limits their ability to precisely map and understand the nuanced multifactorial patterns of intra-regional variation. This study addressed this issue by integrating multiple, complementary spatial statistical techniques. Significant variations in crime patterns also exist in the context of rural-urban dynamics. In urban contexts, highintensity districts such as North 24 Parganas, South 24 Parganas, Kolkata, and Murshidabad record major crimes such as murder, theft, and gender-based violence. Conversely, rural areas experience offences such as theft, arson, and violence against women (Karmakar, 2013; Mukherjee, 2020) [14,16] . The literature consistently highlights the profound mental, physical, and social impacts of these crimes on affected women (Sarkar, 2023) [17] . According to the National Crime Records Bureau (NCRB) [18, 19, 20, 21, 22, 23] , the rate of CAW (IPC) in India surged by over 14.7% between 2017-2022 with the crime rate increasing from 57.9 to 66.4 per 100, 000 women and in case of CAW (IPC+SLL) it increased to 15.1% from 58.7 to 67.6 per 100, 000 women. The most common offences reported were cruelty by husbands or relatives, kidnapping and abduction of women, and assault on women with intent to outrage her modesty, with growth rates of 19.2%, 18.7% and 19.2%, respectively. If we look at the growth rate of CAW (IPC), we can find that states such as Uttar Pradesh (17.40%), Maharashtra (41.8%), Rajasthan (10.60%), West Bengal (12.10%), and Madhya Pradesh (7%) are foremost during the period and these states accounted for more than half (50.20) of the total CAW (IPC) in 2022. West Bengal, in particular, consistently ranked among the top five Indian states in CAW. From 2017 to 2022, cruelty by husband or his relatives, assault on women with intent to outrage her modesty, kidnapping and abduction of women, rape, and dowry deaths were major contributors to CAW. In 2022, the state ranked fourth nationally with 31, 969 cases, leading the national chart in 2021 for cruelty by husbands or his relatives with 19, 952 cases, dropping to second in 2022 with 19, 650 cases. The state also ranked fifth in dowry deaths and kidnapping and abduction of women, seventh in human trafficking, and eleventh in both rape and assault on women. This data paint a grim picture of women's safety in West Bengal, revealing a widespread and escalating crisis. The significant increase in reported cases and the high prevalence of severe offences indicate women's routine vulnerability to violence, often within their homes or at the hands of known individuals. The consistent high rankings of West Bengal underscore that these are not isolated incidents but are systemic issues. Previous research has employed various Geographic Information System (GIS) techniques and statistical methods to analyze the spatial distribution and patterns of CAW and to identify districts with varying levels of crime concentration. However, there is still a lack of precise understanding of the spatial concentration and underlying dynamics of different types of crime. This is largely due to the absence of integrated methodological approaches that combine complementary analytical strengths for a comprehensive and accurate assessment. Therefore, this study aims to query the spatio-temporal patterns and geographical concentrations of CAW, with emphasis on major forms of IPC crimes that occurred against women across the districts of West Bengal between 2017 and 2022, and suggest their policy implications. To this end, the study underscores the necessity of applying the Crime Concentration Index (CCI), Getis-Ord Gi* statistic (Gi* statistic), and Anselin's Local Moran's I (LMI) to achieve the study aims.

Materials and Methods

Data Sources and Acquisition

In this study, district-wise and crime head-wise crime incidence data on CAW (IPC and SLL) were obtained for the years 2017 to 2022 from the NCRB, a recognized, reliable, and authoritative institution that compiles and publishes comprehensive annual crime data. Besides, population data for the year 2011 were acquired from the State Statistical Handbook [24] and projected population data for 2021 were obtained from Pandit and Halder (2023) [25] .

Variables and Indicators

The analysis used four dimensions to measure the overall and gender-specific occurrence of crime. These are Total Crime (IPC+SLL), Total Crime (IPC), CAW (IPC+SLL), and CAW (IPC). Within CAW (IPC), the study considered six major forms of variables such as dowry deaths, cruelty by husband or his relatives, kidnapping and abduction of women, human trafficking, rape, and assault on women with intent to outrage her modesty. The indicators computed across the crime dimensions included district-wise crime rate (per 100, 000 female population), annual growth rate (AGR), compound annual growth rate (CAGR), percentage share, net change (2017-2022), and descriptive statistics such as sum, mean, median, min., max., range, and standard deviation (SD).

Analytical techniques

For each dimension, several indicators were computed to capture temporal and spatial variations in crime. Crime rates were normalized per 100, 000 population using the following standard formula:

Where CR i,j denotes crime rate in the i th year of j th unit of study, denotes total number of reported crimes in the i th year of the j th unit of study, P i,j represents corresponds to the relevant population denominator (i.e. total female population for crime rate against women and total population for overall crime rate) in the same year and unit. The scaling constant 10 5 converts the ratio into cases per 100,000 population to facilitate comparisons across regions and periods.

The AGR was calculated using the following formula:

Where denotes the annual growth rate of crime incidence in the i th year of the j th unit of study, represents the total crime incidence in the i th year of the j th unit of study, and corresponds to the crime incidence in the previous year of the same unit of study. The percentage share was derived to assess the relative burden of CAW (IPC+SLL) against Total Crime (IPC+SLL), CAW (IPC) against Total Crime (IPC), and the six major forms of CAW (IPC) against Total Crime (IPC+SLL). To capture the variability and distributional characteristics, descriptive statistics such as sum, mean, median, min., max., range, and standard deviation (SD) were computed. Additionally, the Net Change was calculated, representing the difference between the first and last years of the study period, and the CAGR was calculated using the following formula:

Where CAGR j = Compound annual growth rate of crime incidents in the j th unit of study, CR f,j = Crime Incidents in the final year ( f ) of the j th unit of study, CR i,j = Crime Incidents in the initial year ( i ) of the j th unit of study, n = Number of years between the initial and final year. Furthermore, districts were classified into seven trend types, such as episodic surge, fluctuating rise or decline, highly volatile, persistent rise, stable or persistent rise, suppressed decline, and suppressed or volatile, based on AGR patterns, standard deviation thresholds, range values, and directionality. To enhance interpretation, heat maps were prepared using the AGR of CAW (IPC+SLL) and major forms of CAW (IPC) for all districts and for West Bengal overall, with class intervals (CI) determined using the following formula:

Where and denote the maximum and minimum growth rates, respectively, in the i th year across all study units, and N represents the total number of study units under consideration (for example, 19 districts of West Bengal). This formulation partitions the observed range of growth rates into equal segments corresponding to the number of study units, thereby ensuring consistent categorization and facilitating comparative analyses across districts and time periods. To analyze the spatial concentration of the six major types of CAW (IPC) across districts, the Crime Concentration Index (CCI) technique was applied to each type of crime. Additionally, a weighted CCI was constructed by integrating these categories into a single composite measure to capture the overall burden of CAW. The CCI was calculated as follows: Where = Crime Concentration Index for the j th district, = Rate of a specific crime against women in the j th district, = Rate of total crimes against women in the j th district, = Rate of the same specific crime against women in the entire state, and = Rate of total crimes against women in the entire state. The CCI measures the geographical concentration of specific crimes within districts and compares local and overall state crime rates. By calculating the ratio of specific crime rates in districts to state-wide rates, the CCI identifies regions with disproportionately high or low crime rates, which helps identify districts that require targeted interventions or further investigation due to unusual crime concentrations. This method provides a nuanced understanding of the distribution of crime. It is grounded in well-established criminological approaches applied by scholars like Brantingham and Brantingham (1997) [26] , McCord and Ratcliffe (2007) [27] , Zhang and Peterson (2007) [28] , Andersen (2009) [29] , Brosnan (2017) [30] , Mondal and Pal (2017) [31] , and Beconyte (2021) [32] provides a nuanced understanding of the distribution of crime. It goes beyond mere counts and contextualizes local crime rates within a broader state framework, thus enhancing the analysis of the spatial crime distribution. To analyze spatial concentration, the CCI values were categorized into five categories using the Natural Breaks (Jenks) method i.e., Very Low Clustering (0.0000 -0.5817), Low Clustering (0.5818 -1.1634), Moderate Clustering (1.1635 -1.7451), High Clustering (1.7452 -2.3268), and Very High Clustering (2.3269 -2.9086). To further identify statistically significant hot spots and cold spots and measure spatial clustering and localized intensity patterns, the CCI (and weighted CCI) was enhanced using the Gi* statistic (1992, 1995) [33] . This combined approach allowed to identify local spatial clusters. where x j is the attribute value for feature j, w i,j is the spatial weight between features i and j, and n is equal to the total number of features. This integrated approach was employed to identify local spatial clusters of high and low crime rates, commonly referred to as hot and cold spots, respectively. While the CCI identifies regions with disproportionately high or low crime rates, the Gi* statistic focuses on local patterns, pinpointing specific areas with significantly high or low crime rates. By capturing these local spatial patterns, the Gi* statistic offers a finer resolution of crime distribution, facilitating targeted interventions. This method refines the analysis by identifying the precise locations of crime hotspots and cold spots, which are crucial for resource allocation and policy-making. To further augment spatial robustness, the CCI was adjusted for spatial autocorrelation by applying Anselin's Local Moran's I (LMI) (1995) [34] . This complementary technique is particularly valuable for detecting spatial clusters and outlier regions, offering insights into the underlying spatial dependencies of the crime data. Incorporating the LMI adds an additional layer of analytical precision. It identifies areas where high or low values are spatially correlated and where anomalies deviate significantly from their spatial neighbours. Mathematically, the index is defined as:

Where I i is the Local Moran's I for location i, x i is the value at location i, x̄ is the mean of variable x, S 2 is the variance of x, w ij is the spatial weight between locations i and j, and x j is the value at location j. This method offers granular insights into spatial dependencies by pinpointing districts that deviate from general patterns and highlighting local clusters. By uncovering localized trends and outliers that broader analyses might miss, the LMI enhances the understanding of the spatial heterogeneity of crime rates. This ensures that areas with unusual patterns are not overlooked, facilitating targeted interventions and in-depth investigation. By employing a combination of CCI, Gi* Statistic, and LMI, this study ensured a comprehensive analysis of crime patterns, local spatial clusters, and outliers. These complementary techniques provide a multifaceted view of crime distribution, leveraging local spatial statistics to capture the full spatial dynamics and deliver accurate, reliable, and nuanced insights into the geographical and temporal patterns of crimes against women in West Bengal.

Statistical tools and Software

Microsoft Excel (Office 365) was used for incidence tabulation, crime rate calculations, and descriptive statistics. ArcGIS 10.8 was used for spatial statistical analysis, including the CCI, Gi* Statistic, and LMI,

Methodological rigor and Limitations

The study framework emphasizes methodological transparency, comparability, robustness, and reproducibility. Each stage from incidence calculation, rate standardization, mean estimation, CCI development to spatial refinement was explicitly documented with formulas to ensure clarity. Crime rates were standardized per 100,000 female population to enable cross-district comparison, and sequential refinement from CCI through Gi* statistic to LMI reduced bias and improved spatial accuracy. Although the cartographic outputs and spatial procedures generated in ArcGIS 10.8 are software dependent, the workflow remains replicable using NCRB data, census projections, and standard spatial statistics tools. However, this study has some limitations. The study relies on 2011 female population data for crime rates for 2017-2020, while projected data for 2021 and 2022 are used due to the unavailability of recent government census data. This reliance on older or projected figures is a limitation, as demographic changes can influence the precision of crime rate calculations. Although it is common practice to use 2011 census data as a previous survey period in the absence of updated census data, this can potentially affect the accuracy.

Results: 4.1 Spatio-temporal trends and patterns of crime against women 4.1.1 Statewide trends and patterns of crime against women (IPC and SLL)

Table 1 and figure 1 illustrate the trends in CAW (IPC+SLL) and overall crime in West Bengal. While the total crime rate (IPC+SLL) declined, CAW (IPC+SLL) remained resilient and volatile, growing in prominence within the state's crime profile. The total crime rate (IPC+SLL) consistently declined from 195,537 cases in 2017 to 180,539 in 2022, with a CAGR of -0.01. Conversely, CAW (IPC+SLL) increased from 30,992 in 2017 to 34,738 in 2022, a net increase of 3,746 cases, with a mean AGR of 2.27%. This increased the proportion of CAW (IPC+SLL) within the total crime from 15.85% in 2017 to 19.24% in 2022 (mean: 17.86%), underscoring the growing prominence of gender-based violence. In addition, the AGR of CAW (IPC+SLL) show significant volatility.

For instance, its declines in 2018 (-1.93%) and 2019 (-1.93%) were followed by a sharp rise in 2020 (20.12%), succeeded by subsequent contractions in 2021 (-1.71%) and 2022 (-3.19%). Besides, IPC crimes also showed a similar trend, the total crime rate (IPC) decreasing from 163,999 in 2017 to 156,503 in 2022 (CAGR: -0.01), while CAW (IPC) cases increased from 28,756 to 31,969. The share of CAW (IPC) within total crime (IPC) rose from 17.53% in 2017 to 20.43% in 2022, averaging 19.62% over the period. Notably, CAW (IPC) sharply escalated in 2020, growing by 30% and elevating its share to 23.05% of total IPC crimes.

District-level trends and patterns of crime against women (IPC and SLL)

The district-level analysis of the AGR for CAW (IPC+SLL) revealed a highly heterogeneous and volatile landscape (table 2 and figure 2). The pattern is predominantly characterized by episodic surge districts, which are distinguished by high positive means and extensive ranges. Malda (mean AGR: 30.30%, range: 95.77, SD: 47.06) is particularly notable for its extreme fluctuations, exhibiting explosive growth in 2017-18 and 2018-19 (80.62%), followed by contractions in subsequent years. Similarly, Purba Medinipur (mean AGR: 9.42%, range: 93.05, SD: 36.85) and Nadia (mean AGR: 6.38%, range: 83.25, SD: 34.81) experienced sharp increases in 2019-20 (74.08% and 68.22%, respectively), followed by Jalpaiguri (mean AGR: 11.41%, range: 62.87, SD: 26.09) and Bardhaman (mean AGR: 6.89%, range: 74.32, SD: 28.15) also reflect episodic surges, with peaks in 2019-20 and subsequent contractions. South 24 Parganas (mean AGR: 15.27%, range: 84.84, SD: 35.82) recorded a dramatic escalation in 2017-18 (50.21%) but declined sharply in 2019-20 (-34.62%). These districts represent zones of instability, where socio-political shocks, enforcement changes, or heightened awareness campaigns may have amplified reporting in certain years. In contrast, several districts exhibited a suppressed decline, characterized by a consistently negative AGR and relatively low volatility. Hooghly (mean AGR: -4.12%, SD: 7.21) and Purulia (mean: -8.24%, SD: 7.29) demonstrated steady contractions, suggesting either genuine reductions or possible under-reporting. Howrah (mean: -6.02%, SD: 13.25) and Murshidabad (mean: -5.64%, SD: 37.02) reflect suppressed but volatile trajectories, with sharp declines interspersed with occasional surges. Cooch Behar (mean AGR: -3.17%, SD: 22.74) also falls into this category, with negative growth in most years but a notable spike in 2019-20 (29.46%). Alongside these, subset of districts exhibited a consistent or stable increases with minimal volatility. Birbhum (mean AGR: 1.78%, SD: 7.58) and Darjeeling (mean AGR: 0.61%, SD: 6.91) demonstrated gradual increases, indicating the structural persistence of genderbased violence without significant fluctuations. Kolkata (mean AGR: 0.85%, SD: 10.90) and Bankura (mean AGR: 3.70%, SD: 17.48) are categorized as experiencing fluctuating increases, characterized by moderate instability but overall positive trajectories. These districts suggest entrenched patterns of violence, where crimes against women remain consistently present, albeit without extreme increases in their numbers. The most extreme cases are highly volatile districts, where AGR values oscillate between sharp declines and increases. Dakshin Dinajpur (mean AGR: 0.52%, SD: 63.02, range: 150.46) recorded both the steepest decline (-50.46% in 2017-18 and 2018-19) and the highest surge (100% in 2019-20). North 24 Parganas (mean AGR: -5.20%, SD: 45.81, range: 105.97) fluctuated between -32.80% and 73.17%. Paschim Medinipur (mean AGR: -1.97%, SD: 28.36, range: 72.81) and Uttar Dinajpur (mean AGR: -2.98%, SD: 24.35, range: 58.62) also fall into the fluctuating decline category, reflecting instability without clear directional trends.

Cruelty by husband or his relatives:

Cruelty against women, encompassing physical, psychological, and emotional abuse by husbands or their relatives, shows significant regional variation (table 7; figure 5). According to the CCI, Hooghly (CCI: 1.2379, moderate) and South 24 Parganas (CCI: 1.2179, moderate) emerged as areas with elevated concentrations. Low clustering levels were observed in most other districts, including Howrah (CCI: 1.1628, low), North 24 Parganas (CCI: 1.0661, low), Nadia (CCI: 1.0120, low), and Kolkata (CCI: 0.8030, low). Dakshin Dinajpur (CCI: 0.5653, Very Low) exhibited the lowest concentration, indicating either minimal reported prevalence or potential underreporting. Spatial autocorrelation analysis indicated more pronounced clustering patterns than those suggested by CCI alone. The statistically significant hot spots included North 24 Parganas (Gi*Z: 2.7190, p<0.01), Nadia (Gi*Z": 2.1584, p<0.05), Howrah (Gi*Z: 2.1629, p<0.05), Hooghly (Gi*Z: 2.1049, p<0.05), South 24 Parganas (Gi*Z: 1.7957, p<0.10), and Kolkata (Gi*Z: 1.7957, p<0.10), indicating spatially concentrated zones. Conversely, Dakshin Dinajpur (Gi*Z: -2.0990, p<0.05), Uttar Dinajpur (Gi*Z: -1.7780, p<0.10), and Malda (Gi*Z: -1.8395, p<0.10) were identified as cold spots, highlighting the socio-spatial divergence of violence. The LMI further validates this finding. High-high clusters were confirmed in North 24 Parganas (LMiZ: 2.6360, p<0.01), Nadia (LMiZ: 2.0832, p<0.05), Howrah (LMiZ: 1.8647, p<0.05), and Hooghly (LMiZ: 1.7122, p<0.05), indicating a statistically significant spatial dependence. Malda (LMiZ: 1.8485, p<0.05) and Uttar Dinajpur (LMiZ: 1.6491, p<0.05) were identified as low-low clusters, suggesting contiguous zones of low but persistent abuse. Kolkata (LMiZ: -2.4075, p<0.01) presents a unique case as a low-high outlier, reflecting internal intensity surrounded by low-crime neighbours. In addition, districts such as Darjeeling (LMiZ: -1.2263) and Murshidabad (LMiZ: 0.2504) exhibited low CCI and nonsignificant clustering, indicating spatial dispersion. Similarly, Birbhum (CCI: 0.8268; Gi*Z: 0.2779; LMiZ: -0.7369) and Paschim Medinipur (CCI: 0.8096; Gi*Z: 1.0188; LMiZ: -1.2993) demonstrated low concentrations and lacked significant spatial dependence.

Kidnapping and abduction of women

The spatial distribution of kidnapping and abduction incidents involving women varied significantly across districts (table 8

Rape

The spatial distribution of rape incidents indicates significant concentrations in specific districts, as shown in table 9 andfigure

Six year mean spatial clustering and concentration of major crimes against women (IPC)

The spatial distribution of the six-year mean of major forms of crimes against women reveals a complex and varied landscape of vulnerability, with districts showing different levels of clustering and spatial dependence (table 10 and figure 8). According to the Mean CCI, none of the 19 districts fell within the very high (> 2.

Discussion

This study presents a comprehensive analysis of the temporal trends and spatial patterns of CAW, offering critical insights into the evolving landscape of gender-based violence. The findings reveal a concerning paradox, where while the overall crime rate (IPC+SLL) has generally declined, the proportion and, in some cases, the absolute number of CAW has increased. Therefore, there is a need for deeper inquiry into the underlying socio-cultural and systemic factors which contribute to this disparity. This may indicate more effective institutional response mechanisms or increased legal awareness. In urban settings, where anonymity is a major issue, such incidents are likely to be significantly underreported, possibly undermining traditional kinship-based violence (Nigam, 2020; Rajat, 2023) [36,37] . Therefore, it is important to interpret these results with caution. The spatial diagnostic analysis of CAW (IPC) using Gi* Statistic identifies concentrated hot spots and underlying spatial dependencies, which are essential for interpreting regional disparities and for taking targeted policy measures to address these issues. Districts such as Uttar Dinajpur, Dakshin Dinajpur, and Malda have consistently emerged as statistically significant hot spots for various crime categories, especially kidnapping and abduction and rape. This observation suggests a spatial concentration, and based on prior research (Islam and Saikia, 2025) [35] , this pattern may indicate a potential socio-cultural reinforcement of violence, although direct evidence for this specific link is beyond the scope of the current study. These findings suggest a pattern that aligns with this theoretical framework, where spatial autocorrelation may indicate systemic tolerance or inadequate resistance to crime. Conversely, districts such as Kolkata and Howrah appear as significant cold spots for dowry deaths and cruelty towards women. However, further qualitative and socio-economic research is required to empirically validate these potential influences. Nonetheless, it is crucial to recognize that low clustering should not be interpreted as an indicator of low prevalence as underreporting and gaps in institutional data in crime statistics are well-established challenges (Kumari et al., 2025) [38] . This is particularly relevant as crimes are severely underreported in both public and private spheres due to fear of social stigma and reputational harm (Mukherjee, 2020; Sarkar, 2023) [16,17] . The identification of recurrent spatial autocorrelation through the LMI highlights that crimes are not isolated incidents but are embedded in regional contexts. The findings emphasize the necessity of multisectoral approaches that integrate legal frameworks, community engagement, and place-based interventions (Rashid & Bhat, 2025) [39] . For example, the spatial outlier observed in Bardhaman necessitates different policy attention owing to its distinct status as a high-risk neighbourhood. Upon thoroughly examining the results, this study identified distinct crime patterns across districts. In particular, dowryrelated deaths are prevalent in Uttar Dinajpur, Dakshin Dinajpur, Bankura, and Bardhaman districts. Incidents of cruelty against women are concentrated in Nadia, North 24 Parganas, Howrah, South 24 Parganas, and Hooghly districts. Crimes related to kidnapping and abduction are prominent in the Malda, Nadia, Uttar Dinajpur, and Dakshin Dinajpur districts. Human trafficking is notable in Darjeeling, North 24 Parganas, Uttar Dinajpur, Jalpaiguri, and Howrah. The occurrence of high rape cases were observed in Dakshin Dinajpur, Murshidabad, Uttar Dinajpur, and Malda. The assaults on women are concentrated in Bankura, Darjeeling, North 24 Parganas, Paschim Medinipur and Purulia. Collectively, the spatial concentration of weighted CAW across these six categories was highest in Purba Medinipur, Bankura, Malda, Jalpaiguri, and Uttar Dinajpur, indicating that these districts face a greater challenge in terms of protecting women from various crimes. The three-pronged approach utilized in this study, which incorporates general clustering, statistically significant hotspots and localized dependencies, provides a comprehensive framework for studying spatial criminology. This multi-level diagnostic method also enhances policy utility by identifying the prevalence of crime in a region and its spatial mechanism. In this context, reference can be made to UN Women (2021) [40] , which states that spatial targeting in violence prevention strategies can significantly improve intervention outcomes, especially when integrated with local governance and community-based safety audits.

Limitations and Future Research:

This study provides valuable insights into the spatiotemporal patterns of crimes against women; however, it is subject to several limitations. First, reliance on secondary data from the NCRB means that the analysis is constrained by the accuracy and completeness of official crime reporting, which is known to be affected by under-reporting due to social stigma and gaps in institutional data. This implies that the identified cold spots might, in some instances, reflect underreporting rather than a genuine absence of crime, potentially skewing the true spatial distribution of crime. Low clustering in certain areas may not genuinely reflect low prevalence but rather significant underreporting, particularly for sensitive crimes such as rape. This means that the findings represent reported crime patterns that may not fully capture the scope of gender-based violence. Second, due to the lack of recent government data, female population data from the 2011 census were used to determine crime rates for 2017-2020. Projected data for 2021 were used for 2021 and 2022. Therefore, reliance on such old and projected data may subtly influence the calculations, representing a minor limitation of this study. Third, this study employs advanced spatial statistics to execute spatio-temporal patterns of reported crimes. While the discussion section offers some theoretical interpretations, it does not incorporate socio-economic covariates (e.g., literacy rates, economic independence, and specific enforcement mechanisms) directly into spatial models. Future research would benefit from integrating such variables directly into spatial models to provide a more comprehensive understanding of the underlying socioeconomic drivers of crime and strengthen the causal interpretations of the spatial patterns identified in this study.

Conclusion, Implications, and Recommendations

The present analysis reveals that the spatial concentration of gendered violence is shaped by complex interactions among geographical factors, social norms, and deficiencies in law enforcement. The findings of this study have several implications for policy and practice. First, the increasing proportion of CAW amidst a general decline in overall crime rates underscores that gender-based violence represents a distinct societal challenge, necessitating focused attention on it. However, this issue cannot be effectively addressed using general crime reduction strategies alone. Second, notable spatio-temporal variations highlight the inadequacy of uniform, state-wide interventions in addressing this problem. Instead, a districtspecific approach, informed by detailed spatial analysis is essential. For instance, districts identified as hot spots for dowry deaths require interventions that differ from those experiencing high rates of human trafficking or domestic violence. The policy recommendations are as follows: Targeted resource allocation: It is essential to allocate additional resources, including law enforcement personnel, specialized units focused on women's safety, and support services such as shelters, counselling, and legal aid to identified hot spots and clusters. Community-based interventions: Implementing localized awareness campaigns, gender sensitization programs, and community policing initiatives in identified hot spots, such as Malda and Uttar Dinajpur for rape and kidnapping, and North 24 Parganas for cruelty, is crucial for addressing the root causes and fostering a safer environment for women. Strengthening reporting mechanisms: Trust in the justice system must be enhanced by ensuring victim support, confidentiality, and swift action on reported cases of sexual violence. Additionally, reporting procedures should be simplified and made more accessible, particularly in areas where underreporting may be prevalent (e.g., human trafficking). Inter-agency collaboration: Strengthening collaboration between the police, judiciary, social welfare departments, NGOs, and local communities is vital to ensure a holistic response to CAW. Data-driven policy making: Continuous monitoring of spatio-temporal trends using sophisticated analytical tools is required to adapt policies and interventions in real time.

In conclusion, this study offers a detailed analysis of crimes against women in West Bengal, highlighting the troubling increase in the proportion of total crimes and significant spatial concentration in specific districts. Identifying particular hot spots and clusters for various forms of CAW provides a practical framework for policymakers to formulate targeted intervention strategies to combat these crimes. By promoting data-driven, localized, and multifaceted interventions, the state can strive to create a safer and more equitable environment for women, thereby effectively addressing this persistent and critical societal issue.

References

1. United Nations (UN). Declaration on the Elimination of Violence against Women. United Nations General Assembly (48th Session: 1993-1994). New York; 1993.
2. Council of Europe. Council of Europe Convention on Preventing and Combating Violence against Women and Domestic Violence (Istanbul Convention). Council of Europe Treaty Series No. 210. Istanbul; 2011.
3. Hampton RL, Gelles RJ. Violence toward Black women in a nationally representative sample of Black families. Journal of Comparative Family Studies. 1994;25(1):105-119.
4. Roy A. Violence Against Women. New Delhi: Rajat Publication; 2003. p. 242.
5. González-Brenes M. Domestic violence and household decision-making: Evidence from East Africa. Unpublished manuscript. Berkeley: University of California; 2004.
6. Rowe WS, Fakih-Sutan N, Dulka IM. A study of domestic violence against academic working wives in Medan. International Social Work. 2006;49(1):41-50.
7. Madhu. Gender differentiation and women empowerment with reference to the plays of Mahesh Dattani. Shodh Sari-An International Multidisciplinary Journal. 2023;2(4):138-151.
8. Lenton RL. Power versus feminist theories of wife abuse. Canadian Journal of Criminology. 1995;37(3):305-330.
9. Chesney-Lind M. Patriarchy, crime, and justice: Feminist criminology in an era of backlash. Feminist Criminology. 2006;1(1):6-26.
10. Niaz U. Violence against women in South Asian countries. Archives of Women's Mental Health. 2003;6(3):173-184.
11. Malhotra A, Vanneman R, Kishor S. Fertility, dimensions of patriarchy, and development in India. Population and Development Review. 1995;21(2):281305.
12. Paul PK. Microfinance, sustainable entrepreneurship, and women's empowerment in the Indian economy. In: Advances in Finance, Accounting, and Economics. IGI Global; 2025. p. 85-110.
13. Lolayekar AP, Desouza S, Mukhopadhyay P. Crimes against women in India: A district-level analysis (19912011). Journal of Interpersonal Violence. 2020;37(910):NP7289-NP7314.
14. Karmakar S. Vulnerable condition of women in South 24 Paraganas district, West Bengal. International Journal of Scientific and Engineering Research. 2013;4(11):641-651.
15. Prasad G. Analysis of women victimization in West Bengal. International Journal of Research in Social Sciences. 2018;8(7):831-835.
16. Mukherjee EB. Spatial variation of different type of crimes in West Bengal. International Journal of Multidisciplinary Research and Development. 2020;7(9):19-30.
17. Sarkar U. A study on crime on women in West Bengal. International Journal of Multidisciplinary Educational Research. 2023;12(2):72-73.
18. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2017.
19. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2018.
20. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2019.
21. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2020.
22. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2021.
23. National Crime Records Bureau (NCRB). Crime in India: Statistics. Vol. I. New Delhi: Ministry of Home Affairs, Government of India; 2022.
24. State Statistical Handbook. Table 1.3: Population and Density of Population Per Square Kilometre by District in West Bengal. Bureau of Applied Economics and Statistics, Department of Statistics and Programme Implementation, Government of West Bengal; 2015.
25. Pandit K, Halder S. Explaining crime against women in the districts of West Bengal: An analytical and empirical analysis. International Journal of Law Management and Humanities. 2023;6(4):111-125.
26. Brantingham PL, Brantingham PJ. Mapping crime for analytic purposes: Location quotients, counts, and rates. In: Crime Mapping and Crime Prevention. 1998;8:263288.
27. McCord ES, Ratcliffe JH. A micro-spatial analysis of the demographic and criminogenic environment of drug markets in Philadelphia. Australian and New Zealand Journal of Criminology. 2007;40(1):43-63.
28. Zhang H, Peterson MP. A spatial analysis of neighbourhood crime in Omaha, Nebraska using alternative measures of crime rates. Internet Journal of Criminology. 2007:1-31.
29. Andresen MA. Measuring crime specializations and concentrations. In: Bruinsma GJN, Weisburd DL, editors. Encyclopedia of Criminology and Criminal Justice. New York: Springer; 2014. p. 3010-3023.
30. Brosnan S. Crime concentration in Ireland in 2012: A location quotient approach. Irish Journal of Applied Social Studies. 2017;17(1):76-93.
31. Mondal T, Pal B. A geographical study of dowry deaths in twenty-first century India. Asian Profile. 2017;45(5):387-403.
32. Beconyte G, Govorov M, Balciunas A, Vasiliauskas D. Spatial distribution of criminal events in Lithuania in 2015-2019. Journal of Maps. 2021;17(1):154-162.
33. Islam JA, Saikia M. Exploring crime against women in India: Socio-economic and spatial dimensions. Society Register. 2025;9(2):51-72.
34. Nigam S. Women and Domestic Violence Law in India: A Quest for Justice. 1st ed. Oxon and New York: Routledge, Taylor and Francis Group; 2020. p. 98-100.
35. Rajat. Challenges and opportunities in implementing feminist legal reforms for gender justice in India. International Journal of Novel Research and Development. 2023;8(2):D1-D9.
36. Kumari S, Kumar N, Sharma A. Mapping the dynamics of crime against women in India: A spatio-temporal analysis. Belgian Journal of Geography (Belgeo). 2025;1:1-21.
37. Rashid S, Bhat YH. Understanding and addressing violence against women: From root causes to effective solutions. The International Journal of Indian Psychology. 2025;13(2):448-458.
38. UN Women. Women's Representation in Local Government: A Global Analysis. Working Paper. Governance and Participation Section, UN Women. New York; 2021.