Rainfall induced landslide is one of the main geological hazard in Italy and in the world. Each year it causes fatalities, casualties and economic and social losses on large populated areas. Accurate shortterm predictions of landslides can be extremely important and useful, in order to both provide local authorities with efficient prediction/early warning and increase the resilience to manage emergencies. There is an extensive literature addressing the problem of computing landslide susceptibility maps (which is a classification problem exploiting a large range of static features) and only few on actual short terms predictions (spatial and temporal). The short-term prediction models are still empirical and obtain unsatisfactory results, also in the identification of the predictors. The new aspects addressed in this paper are: (i) a short-term prediction model (1 day in advance) of landslide based on machine learning, (ii) real time features as good predictors. The introduction of explainable artificial intelligence techniques allowed to understand global and local feature relevance. In order to find the best prediction model, a number of machine learning solutions have been implemented and assessed. The models obtained overcome those of the literature. The validation has been performed in the context of the Metropolitan City of Florence, data from 2013 to 2019. The method based on XGBoost achieved best results, demonstrating that it is the most reliable and robust against false alarms. Finally, we applied explainable artificial intelligence techniques locally and globally to derive a deep understand of the predictive model's outputs and features' relevance, and relationships. The analysis allowed us to identify the best feature for short term predictions and their impact in the local cases and global prediction model. Solutions have been implemented on Snap4City.org infrastructure.

Drastic changes into city road traffic may impact in large portions of the city, then hypothetical scenarios have to be analyzed to identify the best solutions to maintain high quality of city services. In this paper, a solution for unexpected or planned events is proposed and validated with the major focus on traffic flow fields. In order to mitigate the effects on wide area, assessments are needed to evaluate the city changes impact on traffic flow in short time. The proposed solution takes into account static, historical, realtime/dynamic, and forecasting information, with long terms and range of Traffic Flow Reconstructions (multiple simulations, predictions and data transformations) integrated with a specific assessment model to provide support for decision makers. Such a solution dynamically reshapes the road network with many connected critical areas and it automatically computes multiple traffic reconstructions in consecutive time slots, while considering the evolution of traffic flow data according to the related traffic re-distribution at junctions, solving their indeterminacy. Each scenario can be grounded for different road graph solutions, and each solution is evaluated by means of specific indicators taking into account traffic flow criticisms, and topological road graph features. The solution herein presented has been developed into the Snap4City framework for Sii-Mobility national mobility and transport action of the Italian Ministry of Innovation and Research.

Fashion retail has a large and ever-increasing popularity and relevance, allowing customers to buy anytime finding the best offers and providing satisfactory experiences in the shops. Consequently, Customer Relationship Management solutions have been enhanced by means of several technologies to better understand the behaviour and requirements of customers, engaging and influencing them to improve their shopping experience, as well as increasing the retailers’ profitability. Current solutions on marketing provide a too general approach, pushing and suggesting on most cases, the popular or most purchased items, losing the focus on the customer centricity and personality. In this paper, a recommendation system for fashion retail shops is proposed, based on a multi clustering approach of items and users’ profiles in online and on physical stores. The proposed solution relies on mining techniques, allowing to predict the purchase behaviour of newly acquired customers, thus solving the cold start problems which is typical of the systems at the state of the art. The presented work has been developed in the context of Feedback project partially founded by Regione Toscana, and it has been conducted on real retail company Tessilform, Patrizia Pepe mark. The recommendation system has been validated in store, as well as online.

No abstract