Seminar: Applications of data analytics in Smart Cities: Spatio-temporal crime prediction; and Epidemic forecasting based on mobility patterns

The steadily increasing urbanization is causing significant economic and social transformations in urban areas, posing several challenges and raising new issues in city development, public policy, and resource management. However, leveraged by a pervasive and large-scale diffusion of sensing networks in modern cities, huge volumes of geo-referenced urban data are collected every day. Such ever-increasing volumes of urban-related data offers the opportunity to apply data analytics methodologies to discover useful descriptive and predictive models, which can support city managers in tackling the major issues that cities face, including, e.g., urban mobility, air pollution, virus diffusion, traffic flows, crime forecasts, etc.

This talk introduces how data analysis and machine learning techniques can be exploited to design and develop data-driven models as valuable support to inspire and implement smart city applications and services. Then, it presents two real-case studies showing how data analysis methodologies can provide innovative solutions to deal with smart city issues. The first one is an approach for spatio-temporal crime forecasting, based on multi-density clustering and auto-regressive models, to automatically detect crime hotspots in urban areas and to reliably forecast crime trends in each hotspot. The experimental evaluation has been performed on Chicago crime data, showing good accuracy in spatial and temporal crime forecasting over rolling time horizons.

The second one is an approach to discover predictive epidemic models from mobility and infection data. In particular, the algorithm first discovers mobility hotspots and patterns. Then, it detects how urban mobility affects the diffusion of epidemic hotspots, by extracting a regression model for each hotspot. The experimental evaluation has been performed on mobility and COVID-19 data collected in the city of Chicago, to assess the effectiveness of the approach in a real-world scenario.

Presented by Eugenio Cesario, Associate Professor, University of Calabria

Eugenio Cesario is an Associate Professor of Computer Engineering at University of Calabria (Italy). His research interests fall in the broad areas of Data Analytics and Parallel/Distributed Data Mining, and include Urban Computing, Smart Cities, Crime Data Mining, Energy-aware Cloud Computing, Cloud\Grid services architectures, Knowledge Discovery applications. He is a member of the Scientific Board of the Ph.D. in ICT of the University of Calabria. He is also member of the Scalable Computing and Cloud Laboratory (DIMES-UNICAL) and co-founder of DtoK Lab s.r.l., a spin-off of University of Calabria.