This project was conceived after the project “Planning Renewable Energy Communities (REC): the perspective of Municipalities and Small and Medium Enterprises (SME)” (PLANREC), that was as well part of the ERIGrid 2.0 Lab Access programme. In this previous project we developed a tool for the forecast of energy performance and metrics of Renewable Energy Communities. The goal of this new project is to investigate whether it is possible to obtain similar results by relying exclusively on socio-economic data of the REC, excluding historical consumption data of individual households from the necessary prediction data. In most proposed approaches, the short or long-term history of energy consumption is used to predict future energy load demands, but the historical load data needed for this kind of model is not always available. With our work we try to overcome the possible lack of this data and enable accurate REC energy forecasts even when the historical data are not available. Even though the proposed model does not require historical load data to make its predictions, it still needs historical load data for the training phase. We searched for public sources that included both electricity load and socioeconomic data for the training and testing of the model. We found many datasets, but the most complete for our purposes was the IDEAL dataset (IDEAL). We created a database in pgAdmin by reworking the IDEAL dataset for our purposes, in which we included the weather data and the location of the households. It is important to notice that, since we do not have any historical data, our task is a regression problem, not a time-series forecasting problem. We tested some regression models: - RandomForestRegressor; - - ExtraTreesRegressor; XGBoost Regressor. From the grid searches that we conducted on these models, we observed that the best performing model was the ExtraTreesRegressor. An important thing to remember is that even though in the best case the ExtraTrees score was slightly better than the RandomForest and XGBoost best score, the results of the three models were very similar. What really made the difference in the model performance was the introduction of synthetic feature related to the load of the households: - Mean_daily_min_load: average of the daily minimum consumption for each household; - - Change_point_down: indicates whether the household is in stand-by (only basic consumption) or if there are other consumption that can be attributed to the presence of people; Change_point_up: indicates the presence of consumption peak. We validated the performance of the model with R-squared score and the final result was a score of 0.82. We also tested the model with 24-hour forecasts for some specific households, achieving a prediction of the consumption patterns with an acceptable degree of accuracy.