The aim of this thesis is predictive analytics in short-term electricity markets. We want to fill the literature gap on intraday markets, and thus we put particular focus on it. However, the research touches on also the day-ahead and balancing markets, but to a lesser extent. The topics that are considered in the course of the thesis are electricity price forecasting (EPF), especially probabilistic forecasting, intraday order arrival, and power trading optimization. Additionally, we contribute to literature on high-resolution electricity demand estimation, and to statistical software literature providing methods for preprocessing of power-related time series. All these topics are relevant both to scientists and practitioners.