Setting sale prices correctly is of great importance for firms, and the study and forecast of prices time series is therefore a relevant topic not only from a data science perspective but also from an economic and applicative one. In this paper, we examine different techniques to forecast sale prices applied by an Italian food wholesaler, as a step towards the automation of pricing tasks usually taken care by human workforce. We consider ARIMA models and compare them to Prophet, a scalable forecasting tool by Facebook based on a generalized additive model, and to deep learning models exploiting Long Short-Term Memory (LSTM) and Convolutional Neural Networks (CNNs). ARIMA models are frequently used in econometric analyses, providing a good benchmark for the problem under study. Our results indicate that ARIMA models and LSTM neural networks perform similarly for the forecasting task under consideration, while the combination of CNNs and LSTMs attains the best overall accuracy, but requires more time to be tuned. On the contrary, Prophet is quick and easy to use, but considerably less accurate.

Usually, wind energy assessment on a new windfarm is conducted with maximum effort prior to the installation of the turbines by using both numerical and experimental investigations. Yet, often the windfarm performances during operation are not as good as expected. This issue can be investigated with a deep analysis of the operational conditions of the windfarm. The large amount of data collected by the SCADA (Supervisory Control and Data Acquisition) systems installed on the turbines can be very helpful. In the present study, the performances of a windfarm were analysed through the elaboration of the SCADA data from a windfarm in southern Italy; in this site, Sorgenia Green installed nine aerogenerators with a rated power of 2 MW each, on a hilly area with gentle slopes. A systematic approach is proposed to isolate the downtime because of malfunctioning and a manifold investigation is applied to the operational phase: several methods (polar efficiency plot, multidimensional graphical analysis, sectorial power curve and misalignment index, respectively) are suggested and applied for unveiling the sectors where the production output is most affected by the wake interactions. Numerical windflow modelling is performed in the wind rose sectors where underproduction is highlighted by a SCADA data analysis: finally, a SCADA database is employed for testing the goodness of the simulations.