Crime Against Women in India (2001-2023) : A Spatial Analysis using K-Means Clustering Approach

According to the recent data of National Crime Records Bureau (NCRB), 51 cases of crime against women are reported every hour in India. The existing literature and recent studies on gender-based violence have shown that over the years Indian states have followed a pattern of clusters in criminal incidents. This paper studies the panel data from NCRB for twenty-three years (2001-2023) to study and identify specific features and spatial distribution of crime against women across the Indian states. The analysis is conducted on the basis of the top five reported crimes identified as Cruelty against Women by Husband and Relatives, Assault on Women, Kidnapping and Abduction, Rape and Insult to the Modesty of Women. The states with persistently high crime under each crime heads have been identified. Further, Kernel Density Estimation is run in R software which reveals the existence of clustering patterns emerging from the data. The paper further uses the K-Means Clustering method to determine how crime is clustered amongst the Indian states. The optimal value of K from the analysis has been identified as 4 – i.e., four distinct spatial clusters viz. Low Crime, Low to Moderate Crime, Moderate Crime and High Crime have been identified among the Indian states. Uttar Pradesh, Madhya Pradesh, Rajasthan, West Bengal and Maharashtra exhibit high crime rates. Further, all north-eastern states are in Cluster 1 showing persistent low crime rates except for Assam which shows up in Cluster 2. The majority of southern states like Tamil Nadu, Kerala, Karnataka, Telangana shows up in Cluster 2 with an exception in Andhra Pradesh being the only state showing up in Cluster 3.