Mobile devices are increasingly becoming a proxy to habits and daily patterns of users. This is opening new possibilities in the area of human behaviour analysis, where the many sensors available on smartphones can be used to regularly monitor users' daily activities and interactions in an non-obtrusive way. In this paper we focus on proximity detection, which refers to the ability of a system to recognize the co-presence of two or more individuals. We present Comm2Sense, our mobile platform to detect proximity among users exploiting sensing capabilities available in modern smartphones, namely Wi-Fi hotspot and Wi-Fi receiver. The platform estimates the distance between subjects applying data mining techniques to the analysis of the Wi-Fi RSSI. We describe the design and implementation of the platform, together with the technical solutions implemented in each module. We demonstrate that the proposed platform is able to achieve a resolution of 0,5 meters.