In recent years, there has been an increas-ing awareness to traffic localization techniques driven by the emergence of heterogeneous networks (HetNet) with small cells deployment and the green networks. The localization of hotspot data traffic with a very high accuracy is indeed of great interest to know where the small cells should be deployed and how can be managed for sleep mode concept. In this paper, we propose a new traffic localization technique based on the combination of different key performance indicators (KPI) extracted from the operation and maintenance center (OMC). The proposed localization algorithm is composed with five main steps; each one corresponds to the determination of traffic weight per area using only one KPI. These KPIs are Timing Advance (TA), Angle of Arrival (AoA), Neighbor cell level, the load of each cell and the Harmonic mean throughput (HMT) versus the Arithmetic mean throughput (AMT). The five KPIs are finally combined by a function taking as variables the values computed from the five steps. By mixing such KPIs, we show that it is possible to lessen significantly the errors of localization in a high precision attaining small cell dimensions.