The wind-structure interaction for structures sensitive to the wind action is commonly investigated through experimental campaigns in wind tunnel. Aerodynamic tests on rigid body models are carried out to estimate pressure coefficients that in turn are used to estimate the wind action on the prototype surfaces. Pressure taps are located on the model surfaces and pressure scanners acquire pressure time histories. Pressure coefficients are calculated from pressure series, they are dimensionless values and they are nondimensionalized on the wind velocity. In the wind tunnel, the flow is assumed stationary and consequently the mean value of pressure coefficient time history is independent of the experiment time length. However, the acquisition time length affects peak values and in particular the peak factor. In addition, the most of processes acquired near edges and near the flow detachment zones are non-Gaussian, contrary to what codes and standard suggest. It was estimated that several processes on the surfaces of a rigid body are non-Gaussian and, in this case, the mean value is not significant. At contrary, the mean value is mostly assumed as representative value of a measurement and it is used to size the structure. It is true only for some specific zones. In addition, they depend on the wind angle of attack. The purpose of this study is to discusses the Gaussian and non-Gaussian processes localization for two examples of structure families, large span roofs and bridge closed box section using aerodynamic experiments given by literature. The geometries subject of this investigation are typical geometries of roofs and bridge sections sensitive to wind action as for example cable nets roofs and suspended bridges.