With the growth of urban cities, more and more energy (electricity) is need for daily human activities, such as cooking, heating, watching TV, charging mobile devices. Hence energy consumption and demand estimation is a key activity to plan energetic matrix, upgrade electricity grid, save money, reduce gas emissions, reduce air pollution and even save water. Energy demand estimation and prediction is not a trivial task for energy providers due to different factors linked to local aspects. For instance, in developing countries urban energy needs and special projects like factories or technological centers are analyzed in separately. While urban needs forecast is correlated with the projected GDP of each country, special projects follow ad-hoc energy demand studies. As for developed countries, urban energy demand and special-purpose project demand are seen together since development is planned in advance. In order to deal with these issues, this collaboration project aims to develop new prediction models to estimate population density so as to classify human activities and to assess energy (electricity) consumption in urban centers using as primary data sources telecommunication, geo-spatial, social media and text data. The overall research will be based on the exploitation of novel Big Data analytics, Machine learning, Social Network Analysis, Data and Text Mining methods and technologies. Specifically, the project will contribute toward research and development of novel Big Data Analytics approaches to develop a model for predicting energy (electricity) consumption based on new computational techniques for tasks as follows: ? Spatio-temporal analysis: to detect activity zones, which are points of interest in a city. ? Mobility model analysis to estimate dynamic density of people in activity zones over time. ? Automatic text analysis: to infer categories (urban, commerce, industrial, etc.) of activity zones based on social network messages.