*5.1. Data Description*

The experimental evaluation presented in this section is performed on the 'Crimes—2001 to present' dataset, consisting of a collection of crime events that occurred in Chicago from January 2001 to the present. The dataset is publicly available on the Chicago Data Portal (https://data.cityofchicago.org/, accessed on 18 December 2022), which also collects and provides open data about various aspects and events of Chicago, e.g., food inspection, traffic crashes, and COVID-19 vaccine diffusion. Each crime in the dataset is both geo-localized (with latitude and longitude) and time-stamped. Furthermore, it includes attributes describing other characteristics of each crime event, e.g., the FBI code and the crime type.

For the sake of our experimental evaluation, we consider only the latitude and longitude of crime events that occurred in 2012 and localized inside the boundary box shown in Figure 10a,b. The area has a perimeter of about 52 km and extends on approximately 135 km. The total number of crime instances is 100,219. The area includes different zones of the city, such as residential, commercial, tourist, and cultural zones, each one characterized by different crime densities. Given such a property, detecting urban hotspots in the area is a good benchmark to compare the performance results of the selected algorithms.

**Figure 10.** Selected area of Chicago and geo-localized crime events. (**a**) Polygon of the area; (**b**) geolocalized crime events.
