Sometimes many PhotoVoltaic (PV) plants have to be monitored by an unique supervision centre. In these cases the real-time analysis of the whole population of data by means of statistical approach requires long time and it is not always possible. After an introduction on Big Data and statistical approaches, the paper presents a procedure of analysis based on sampled data and on the whole population, only for inefficient PV plants. The paper proposes a resampling method, so-called bootstrap, able to approximately describe the sampling distribution. Its peculiarity consists in taking information about the population of the data (whichever its distribution is) and giving quickly preliminary information about the operation of the PV plants.