The rate of citizen calls for police service exhibits marked seasonal
fluctuations, generally reaching a peak during the summer months
and slacking off to a low point in midwinter. Data from St. Louis,
Chicago, and Detroit indicates a high correlation between the
number of calls per week and meteorological variables such as aver
age temperature and hours of daylight. Simple regression models,
using the weekly norms for the weather variables, are employed to
forecast the weekly demand for police service in each of the three
cities. Since the weather data are easily obtained from the local
weather bureau, and only a single year of police data is needed to
estimate the regression coefficients, police planners may find such
climatological models handy for scheduling patrol deployments and
officers' vacations.
