A Multi-indicator Approach for Smart Security Policy Making

Measuring urban crime is a well-established practice for many police forces, local governments and public administrations all over the world. There seems to be, however, a large gap between the rigorous academic definition of how crime indicators should be calculated and interpreted and the actual use that is made of them. Crime counts and rates are unadvisedly put out with little effort to clarify their different meaning; population-based crime rates remain a standard measure despite quite compelling arguments against the use of population as an offset for cross-sectional comparisons; most importantly, little attention seems to have been paid to the consequences of formulating preventive policies based on poorly defined and understood indicators. This issue looks even more evident with high-definition indicators that detail the levels of crime for very small statistical units (streets, street segments and blocks) with situational interventions in mind. After a review of the literature, we illustrate, through a case study, the different landscapes of urban safety and risk of crime when five different families of indicators are alternatively used: crime counts, population-based crime rates, risk-based crime rates, crime density and location quotients. We propose a multi-indicator approach to the ranking and prioritization of urban security issues based on partial order scalogram analysis by coordinates that presents substantial advantages as an operational tool for Public Administrations in a Smart Cities framework.