Varimax Rotation-Based Exploratory Factor Analysis of Crime Dataset

We investigate the use of Exploratory Factor Analysis (EFA) on a crime dataset featuring the number of occurrences of the crime types (like Burglary, Robbery, Theft, etc.) in different location types (like Apartment, Restaurant, Street, etc.). We conduct Varimax rotation of the Eigenvectors of the retained principal components to maximize the communality score for any crime type and show EFA to be helpful in identifying the minimum number of hidden factors to dominantly represent the variations in the number of occurrences of the crime types at the different locations. We extract a mapping of any crime type to a dominating factor (the factor incurring the largest of the factor loadings under Varimax rotation). We observe the crime types mapped to the same dominant factor (different dominating factors) to mostly exhibit stronger positive or strong negative correlation (weak-moderate correlation) amongst themselves with respect to their number of occurrences at different locations.