The proposed project aims at exploiting the forthcoming massive deployment of Internet of Things (IoT) smart metering devices for energy and water markets to develop novel applications for demand management, and environmental and ecological monitoring.The idea is to apply advanced machine learning techniques to the data collected by the smart meters to create detailed consumer profiles, to disaggregate the end-use of the resources, to automatically identify leakage or waste in the distribution networks, and to discover trends and identify positive and negative behaviors on the end consumers.With data insight, tools will be developed for resource management and planning, along with tools and strategies raise awareness on efficient usage.
With the advancements in IoT devices, data transmission and storage we are generating massive amounts of valuable data for diverse applications.The proposed project consists of the study of novel architectures, methodologies and algorithms to analyze the data collected from IoT smart metering devices for resources such as energy and water. The objective is to generate algorithms and tools to help, on one side, utility companies to make accurate demand management and reduce the waste on their networks; and on the other enable end-consumers to track and understand their consumption habits, raise awareness of the importance of efficient consumption, and provide information to correct negative behaviors.The main challenges on the project is to research novel approaches to unsolved problems such as real-time error detection and correction for IoT data transmission since clean data readings are critical for getting unbiased results from machine learning algorithms; and disaggregation of end usage of resources in nonintrusive ways, as retrieving electricity and water consumption at individual appliance level is essential to assess the contribution of different end uses to the total consumption at household or building level.The envisioned research plan requires the integration of multiple fields, such as:
- Machine and Deep learning: Neural networks and machine learning algorithms are powerful tools for addressing problems such as end-usage disaggregation, user profiling and consumption prediction and forecasting, and have achieved significant improvements in the recent years thanks to the support of the research community and the constant grow in computational power available, examples of the application of such techniques and its results can be seen in the works by Cominola et al  and Piga et al .
- Internet of Things: As mentioned before, the main sources of data for the proposed project are smart meters, sensors and data collection infrastructure, which transfers the information through specific IoT protocols and formats. Working with such devices poses challenges on the data transmission, the verification of the data quality and the error detection and correction in real time or during the post-processing phase, in addition to privacy related issues.
- Big Data management: Collecting consumption readings data and metadata at building, household and appliance level implies technical challenges in the management of data, e.g., storing streams of big data sets from various sources in real.
- Gamification of consumption and behavioral data: The application of game mechanisms is a commonly used approach, which is highly effective to motivate users to get involved on certain activities. The design of the most appropriate data visualization and gamification mechanisms will be a significant research question. The combination of gamification and resource usage data has been experimented in recent years, but still there is important work to do to derive effective strategies for improving user engagement towards sustainability topics.
- Environmental and ecological monitoring: User generated content (UGC) and feedback can be used for environmental and ecological monitoring and reporting purposes. Acquiring, analyzing and processing consumer response to water quality or electrical voltage fluctuations can lead to the detection of anomalies of potential interest such as climate change effects, droughts, electric storms, etc., or errors on the distribution network (e.g., leakage, short circuits, etc.) leading to waste. User feedback on resource quality and availability could complement data collected with smart metering infrastructures, improving quality of analysis and prediction of ecological phenomena. Example of this approach applied to water consumption and social networks is the work by Giuliani, M. & Mossina, J. .
- Information visualization: the most appropriate visualization methods depend on the semantics and the purpose of the visualization as well as on the end-user that will interact with them. Dashboards are a powerful tool for expert users looking for data insight and monitoring of indicator variables, while simple yet meaningful interfaces are preferred for general public with basic understanding of information background. For this reason, visualization design must balance information content and visual components.
- Resource demand management: This research area has grown in the last 30 years as the urban areas began to grow and with them the demand of resources such as water, electricity and gas. There is still lot of work in this field, as currently there are only few integrative approaches covering the challenges of deploying a large-scale smart metering infrastructure, of the information system required to store the data collected (ensuring consumer’s privacy) and of the exploitation of the data to derive consumption patterns and trends to ensure an effective planning and forecasting.
 Cominola, A., Giuliani, M., Piga, D., Castelletti, A., & Rizzoli, A. E. (2017). A hybrid signature-based iterative disaggregation algorithm for non-intrusive load monitoring. Applied energy, 185, 331-344.
 Piga, D., Cominola, A., Giuliani, M., Castelletti, A., & Rizzoli, A. E. (2016). Sparse optimization for automated energy end use disaggregation. IEEE Transactions on Control Systems Technology, 24(3), 1044-1051.
 Giuliani, M., & Mossina, J. (2016, April). Reducing the Intrusiveness of Energy and Water End-Use Disaggregation via Social Media and Users Interactions. In EcoMo@ ICWSM.