Toward ECListener: An Unsurpervised Intelligent System to Monitor Energy Communities

Abstract This work describes the architecture of the “Obserbot Cluster”. Obserbot is the name of a long term project resulting from the crasis of the terms “observe” and “bot” (abridged form for robot). The Obserbot Cluster is formed by a number of different components running as micro-services on an dependable and scalable platform. Obserbot main objective is to monitor the web social media to extract some knowledge without any human supervision. The service is available 24 h a day and 365 days a year. Obserbot targets are presently two: official news media and twitter. A rough mass of textual data are collected form those two sources. In-line semantics and sentiment analysis of the collected stream allows to extract information on a specific Domain of Interest (DoI). In particular, Natural Language Processing and machine learning techniques are extensively employed for recognizing named entities, performing events classification and building up taxonomies. The hardware and software architecture enabling such a collection is rather versatile and it can be exploited to accomplish several different purposes. However, the semantic analytics is strictly DoI dependent. In its present form, Obserbot is able to handle information related to essential services and monitor publications on emerging Energy Communities. Essential services are sustained by networks of utilities providing basic goods: water supply, energy supply (e.g. gas and electricity), fuel supply, fresh food supply etc. and responding to other fundamental human needs such as transports, mobility and social connectivity.