A tested method for assessing and predicting weather-crime associations

Few studies have focused on haze as a weather element and its correlation with crime. In this study, we examined haze as a weather variable to investigate its effects on criminal activity. We used both monthly crime data and weather records to build a regression model that contains a sequential statistical approach to reach the correlation coefficients between the variables. Also, we developed a prediction model to predict crime cases considering three weather factors: temperature, humidity, and haze. We applied this model in two different climate provinces in Saudi Arabia, namely, Riyadh and Makkah. Riyadh is a desert area and observes haze approximately 17 days per month on average, while Makkah is a coastal area observing haze an average of 4 days per month. We found a measurable relationship between each of these three variables and criminal activity. We found that a one-degree increase in temperature was associated with an increase in assault of 0.739, when humidity and haze were held constant. For other independent variables measured against the same crime in Riyadh, a one-degree increase in humidity was associated with a 0.164 increase in assault. An increase in the number of times that a haze phenomenon is observed was associated with an increase of 0.359 in assault cases. Haze had the most effect on theft, drug, and assault crimes in Riyadh compared to the other elements. Temperature and humidity have a significant relationship with crime in Makkah, while haze had no significant influence in that region.