The theme of my thesis is the short, medium and long term energy load forecasting, based on weather forecasting data. Since the electricity became widely used, energy load forecasting has become an important problem for every stakeholders. There is a significant business need for accurate forecasts, because of the optimal use of resources, and to maintain the demand supply balance. This is important for both the production industry and the consumers. In the first part of my thesis, I introduce the relevant literature about the prediction methods and evaluation techniques. I describe especially the ARIMA models, regression analysis and artificial neural networks. Thereafter I test the introduced models and I evaluate and compare the various forecasts, to minimize the forecasting error, and to find the proper model. The required energy load data is collected from the ENTSO-E with web scraping. The weather forecast data is from the ECMWF organization’s data centre. This portal has a usable WEB and GRIB api, developed for the purpose of helping the data downloading. I focus on feature creation, to find more and more useful variables. The comparison of plan and fact data series, the findings, and the evaluations are visualized in a meaningful way. The implementation, like data downloading, feature creation, modelling, is with Python programming language.