Human mobility gives rise to the appearance of different face to face interaction networks. These networks are the substrate for many social phenomena occurring on them, such as the competition of opinions, spread of fads, innovations and even diseases. We study a simple agent based model of opinion competition named the Voter Model on networks following characteristics of human mobility. We assess the impact of such mobility patterns on the diffusion of opinions. Recent availability of huge databases on human activities has revealed that the mobility patterns are very recurrent. In particular a very important part of these patterns is governed by commuting activity worldwide. We use data from the census of the US for the commuting patterns and compare also with other proposed models for commuting, such as the gravity law and the recently proposed radiation model. The substrate for the opinion model is thus given by the cities in the country and the commuting fluxes between them. The opinion model is based on the competition of two equivalent opinions. Each agent can have one of two equivalent opinions and simply copy the opinion of another agent she comes in touch with. We measure the impact of commuting mobility on this kind of diffusive random copying behavior.