We report on the new data mining approach based on Random walk theory. Through the full analysis of complex impedance spectra including electrode polarization effect, we derived the number of moving ions and the diffusion coefficient. These useful numbers were extracted without using an equivalent electric circuit. Our approach has the potential to find a widespread use in various electrochemical fields, such as in batteries, solar cells, fuel cells and semiconductor industry in general. In this paper, we show results for a case example of an ionic conductor -Ag-doped chalcogenide glass. However, this approach can be exploited for any semiconducting or ionic conductor materials, either in amorphous or crystalline form.