From smart firms to smart consumers: Complex Systems and Machine Learning for Industry 4.0

The Fourth Industrial Revolution, also indicated as Industry 4.0, has reinvented the way firms design, produce and distribute their products. Technologies such as Industrial Internet of Things (IIoT), cloud connectivity and Machine Learning are now deeply intertwined into the production process. This unified and integrated approach to manufacturing results in products, factories, and assets that are connected and intelligent. Accordingly, a firm can be seen as a Complex System: every aspect of the firm’s activity is strictly linked to each other and their evolution, as well as that of the whole system, depends on these connections. In other words, the evolution of every single element cannot be studied on its own but must be placed in a more holistic framework taking into account its relations with all the other elements. These systems are quantitatively studied through graph theory, which denotes the connected elements as nodes, the corresponding connections as links and allows to define metrics that quantitatively measure different kinds of importance of the nodes in the system. Another key aspect in Industry 4.0 is the possibility of extracting value from data, coming them from Complex Systems or not. These insights can be found through Machine Learning techniques, both in its supervised and unsupervised form. Accordingly, one important example of Complex System in the Industry 4.0 context will be analyzed from different opints of view: the startup ecosystem. This is defined as the set of startups and their funders. An accurate modeling of this system through graphs will be of fundamental importance for its quantitative analysis aiming at highlighting the most important elements in the ecosystem. Moreover, these quantitative information will be deployed in Machine Learning algorithms in order to forecast the most successful startups based on the information given by graph modeling. Moreover, since consumers play a more active and central role in the Fourth Industrial Revolution, a more comprehensive study of their needs and tastes is pivotal for the success of a firm. Among the economic activities, tourism is the most affected by consumers' reviews and it is also one of the most profitable. Accordingly, Natural Language Processing techniques, together with Machine Learning algorithms and Explainability tools will be employed in order to analyze tourists reviews about accomodation facilities in Puglia, a region in the South-East Italy, which has witnessed an explosion of tourists arrival. This analysis reveals what aspects of the Apulian tourist offer are appreciated the most and which have to be improved.