Exploring Urban Fear of Crime using Unsupervised Classification - The Case of Athens

Abstract

This publication examines whether urban landscapes, exhibiting different spatial and thematic characteristics, can affect the fear local residents feel about crime. The purpose of this research is therefore to explore the correlation of the fear of crime in modern cities with urban, environmental and social characteristics. In order to verify this postulation, the Municipality of Athens was set as study area and a wide questionnaire-based survey was conducted. A dedicated web platform (http://www.fearofcrime.com/) was employed to collect, analyse and visualize crowd fear experiences and their spatial distribution. Over 900 participants have contributed relevant data to be further explored. In order to meet the above objective and facilitate the process, the respondents and their answers were grouped into clusters via unsupervised classification. The variables were formed by all respondents’ answers and the k-means algorithm was used. From this process, groupings of a certain number of clusters emerged. The distribution of each cluster was determined on the seven City Districts of the Municipality of Athens, which were used as spatial reference units, revealing the largest and smallest concentrations. Thus emerged the image of how the phenomenon of urban fear of crime is shaped in the different districts of the Municipality of Athens.