**3. SSK methodology**

Our dataset consists of CDRs recorded by an Israeli telecommunications company during a 62-day period, each day between 4 a.m. and 10 p.m., in a region covering a major part of Israel's center district, including the city of Tel Aviv and its neighboring cities. The data include a diverse collection of human activity—a variety of settlements (cities and villages), open areas, highways, and industrial areas.

Our workflow (**Figure 1**) can be divided into five steps: (3.1) Area selection, (3.2) Division of smaller units of land with grid-like partitioning, (3.3) Land-use labeling, (3.4) Feature extraction, and (3.5) Usage of the SSK algorithm for land-use classification.
