The elasticity and structure of graphite nanoplatelets networks in polymer matrices are studied through linear viscoelastic analyses. GNPs-filled polystyrene nanocomposites at different filler content are prepared through a combination of solution and melt mixing techniques. Electrical volume conductivity experiments prove that a continuous path of conductive nanoplatelets builds up across the matrix above a critical filler content. GNP networks, however, are too tenuous to be detected through conventional dynamic-mechanical spectroscopy in the melt state. Nevertheless, we are able to estimate their elasticity by exploiting the predictive feature of a simple two-phase rheological model. Our approach, validated through the building of a master curve of the elastic modulus of samples at different composition, allows to isolate the elastic contribution of the bare GNP network, whose dynamics reveal that its elasticity follows critical behaviour as predicted by percolation theory. (c) 2012 Elsevier Ltd. All rights reserved.