Innovative Data Analytics Systems and Business Applications for Smart Cities and Factories

In the last years, data analytics has become an essential business process for any company; and smart services started to have a crucial role in the market for ensuring smooth business processes. Executives require in-depth knowledge of their business, the market, competitors and customers to make the right decisions and steer their business towards prosperity. In this thesis; we proposed innovative approaches for multiple business sectors which showed brilliant results after testing them. Basically, we focused on analysis applications in two main business sectors: energy and telecommunication. However, from technical point of view, the vast amounts of data usually consume huge storage space on the cloud based servers. Furthermore, cloud-computing started to struggle in satisfying the required quality of ser vices, due to some technical limitations. Towards overcoming the previous vulnerabilities, edge computing is used recently for decreasing latency between processing data center and end-user, balancing network traffic, avoiding network bottleneck and reducing the response time for time-sensitive applications and real-time analysis insights. As an application in the energy sector, we proposed in this thesis an innovative predictive maintenance application which successfully anticipated a fault in the equipment with an advance of almost two weeks, and also it demonstrated its robustness to false alarms during normal conditions in the production line of Enel Green Power in Catania. Moreover, two innovate analysis based application are proposed for improving customer experience and increasing clients’ satis faction about the provided services in the telecommunication sector. Later, a comprehensive survey discussing deeply the recent trends of edge computing is presented and finally, two experiments cloud-to-edge are mentioned comparing results’ reliability while running the same application on multiple edge servers with different characteristics in addition to com pare results’ quality between edge computing and cloud computing. The previous proposed systems were validated through specific experiments