Random networks of silver nano wires have been considered for use in transparent conductive films as an alternative to Indium Tin Oxide (ITO), which is unsuitable for flexible devices. However, the random distribution of nano wires makes such conductive films non-uniform. As electrical conductivity is achieved through a percolation process, understanding the scale-dependency of the macroscopic properties (like electrical conductivity) and the exact efficiency of the network (the proportion of nano wires that participate in electrical conduction) is essential for optimizing the design. In this paper, we propose a computational method for identifying the representative volume element (RVE) of nano wire networks. This defines the minimum pixel size in devices using such transparent electrodes. The RVE is used to compute the macroscopic properties of films and to quantify the electrically conducting efficiency of networks. Then, the sheet resistance and transparency of networks are calculated based on the predicted RVEs, in order to analyze the effects of nano wire networks on the electrical and optical properties of conductive films. The results presented in this paper provide insights that help optimizing random nano wire networks in transparent conductive films for achieving better efficiencies.