Wireless Sensor Networks (WSNs) have recently achieved tremendous success at both research and industry levels. WSNs are currently implemented in many areas, such as the military, environmental monitoring, and medicine. WSN nodes are battery-operated, and energy saving is critical for their survival. Several research papers have been published on how to optimize power usage. In this paper, we focus on improving power consumption by optimizing data transfer. We propose an Energy-Efficient Partitioning Algorithm to reduce data transfer and consequently improve power consumption. Using data collected from a real WSN in the City of Moncton, we implemented and compared the performance of the proposed algorithm with another data reduction algorithm. Experimental results show that our algorithm outperforms a recent data reduction technique in terms of power saving.