Using social media to explore spatio-temporal pattern of the fear of crime

Abstract

The analysis of spatial information and temporal patterns via social media content such as Twitter has received substantial attention recently. Aiming to generate deeper understanding of human sentiment, thousands of tweets including non-georeferenced and georeferenced data from February to May 2021 in England were collected to analyse the behaviour of the public. The novelty of this research is that we integrated social media data analytics and GIS based spatial analysis to detect the effect of homicide over time and space, along with employing χ -expectation surface to further investigate the areas where crime rates are underestimated compared with the English Index of Deprivation (EID). The results of this study reveal that fear, vulnerable emotions and Sarah Everard’s homicide are positively correlated. People tweeted more about the homicide and expressed more fear or vulnerability in the two weeks surrounding the homicide. Regarding the GIS-based spatial analysis, spatial clusters concentrated mostly in London, Manchester, Birmingham which are as expected as these are developed areas. Furthermore, after comparing the spatial data of the homicide and fear emotions with the EID, it showed that the distribution of most of the data was presented as expected as the deprivation data provided by the English Government. This finding can serve as input for further related investigations.