Theory and projects – The SmartData@PoliTO focuses on both fundamentals and applications of data science.

Considering fundamentals, two main research lines are identified:

  1. Algorithms and methodologies for data analysis, with interest on big data processing, data mining, deep learning, and machine learning. Scalability is a major issue, and all these methodologies need to be revisited toward big data, where the Volume, Veracity and Velocity challenge them. New algorithms are needed, capable of exploring datasets characterized by millions of features, intrinsic data sparsity, and different data densities. When data is not enough, simulations may be leveraged to generate datasets to train machine learning models, which in turn can be used to simplify the simulation model, forming a cycle. The center will study the fundamental theory behind such an approach, to understand if and when this process leads to success.
  2. Innovative business model generation: Extracting Value from big data is a top challenge. Taxonomies, business models, business practices, and impact of entrepreneurship, regulations, and policy makers – all must be judged to understand how to create economic value through big data. The center is a reference to provide blue print for managers, policy makers, and educators to evolve business practice, models and education. 

Considering applications, we have identified the following areas:

  1. Predictive maintenance: Historical data allow the definition of models to detect failures in advance, and implement appropriate strategies to reduce maintenance operations. Big data plays a key role, given the system complexity makes it almost impossible to use phenomenological models.
  2. Internet & Cybersecurity: Internet is the biggest source of big data, and is constantly generating cyberthreats, with IoT scaling the challenge to humongous sizes. Network management and security involves the study of traffic, with anomaly detection algorithms applied for cybersecurity threats, detection and countermeasures design.
  3. Mobility analysis: The heterogeneity of user habits, taste and social interactions make the understanding of customers’ needs a challenge. Analysis and modeling of data coming from different platforms is key to provide new offerings, improve customer satisfaction, and system processes.


The following competences are available at SmartData@PoliTO:

  1. Algorithms and methodologies for data analysis (big data processing, data mining, supervised and unsupervised machine learning, deep learning, rule mining, classification, prediction, anomaly detection, clustering, …)
  2. Methodologies for data modeling based on statistics, computational topology, geometry with application to data analysis, graph modeling and information mining from graphs
  3. Study the transformations in industrial structures (value chains, business models, managerial practices)