Clean air in cities improves our health and overall quality of life and helps fight climate change and preserve our environment. High-resolution measures of pollutants’ concentrations can support the identification of urban areas with poor air quality and raise citizens’ awareness while encouraging more sustainable behaviors. Recent advances in Internet of Things (IoT) technology have led to extensive use of low-cost air quality sensors for hyper-local air quality monitoring. As a result, public administrations and citizens increasingly rely on information obtained from sensors to make decisions in their daily lives and mitigate pollution effects. Unfortunately, in most sensing applications, sensors are known to be error-prone. Thanks to Artificial Intelligence (AI) technologies, it is possible to devise computationally efficient methods that can automatically pinpoint anomalies in those data streams in real time. In order to enhance the reliability of air quality sensing applications, we believe that it is highly important to set up a data-cleaning process. In this work, we propose AIrSense, a novel AI-based framework for obtaining reliable pollutant concentrations from raw data collected by a network of low-cost sensors. It enacts an anomaly detection and repairing procedure on raw measurements before applying the calibration model, which converts raw measurements to concentration measurements of gasses. There are very few studies of anomaly detection in raw air quality sensor data (millivolts). Our approach is the first that proposes to detect and repair anomalies in raw data before they are calibrated by considering the temporal sequence of the measurements and the correlations between different sensor features. If at least some previous measurements are available and not anomalous, it trains a model and uses the prediction to repair the observations; otherwise, it exploits the previous observation. Firstly, a majority voting system based on three different algorithms detects anomalies in raw data. Then, anomalies are repaired to avoid missing values in the measurement time series. In the end, the calibration model provides the pollutant concentrations. Experiments conducted on a real dataset of 12,000 observations produced by 12 low-cost sensors demonstrated the importance of the data-cleaning process in improving calibration algorithms’ performances.

Text categorization remains a formidable challenge in information retrieval, requiring effective strategies, especially when applied to low-resource languages such as Italian. This paper delves into the intricacies of categorizing Italian news articles, addressing the complexities arising from the language’s unique structure and writing style. The implemented methodology involves preprocessing the text, generating word embeddings, conducting feature engineering to extract meaningful representations, and training a classifier using the document vectors. The evaluation of the model’s performance is done on a partitioned dataset with a training set for model training and a test set for categorization, allowing assessment of its efficacy on unseen data. Within this paper, we assessed fifteen classifiers for the categorization of Italian news articles, scrutinizing eight models and three approaches for combining word embeddings to derive document vectors. We conducted a comparative analysis between established models such as Word2Vec and FastText and six novel Italian models pre-trained on native datasets. A significant highlight of our work is the introduction of an Italian GloVe model, previously absent for the Italian language. The datasets selected for testing the models’ performances are DICE, a dataset of 10,395 crime news articles extracted from an Italian newspaper, and RCV2-it, a collection of 28,405 Italian news stories released by the multinational media company Reuters Ltd. The tests conducted achieved as the best F-scores 84% and 93%. The results underscore the efficacy of the Support Vector Classification algorithm, while also revealing the inefficacy of Gaussian Naive Bayes, Bernoulli Naive Bayes, and Decision Tree models within the domain of text categorization. The comparison of the word embedding models revealed the better performance of Word2Vec and GloVe concerning FastText. The broader impact of this paper lies not only in advancing text categorization methodologies for Italian documents but also in enriching the linguistic landscape by releasing six novel Italian word embedding models.

One of the main issues in today's cities is related to public safety, which can be improved by implementing a systematic analysis for identifying and analyzing patterns and trends in crime also called crime mapping. Mapping crime allows police analysts to identify crime hot spots, moreover it increases public confidence and citizen engagement and promotes transparency.This paper is focused on analyzing and mapping thefts through on-line newspaper using text mining techniques for an Italian city.

Modern cities face pressing problems with transportation systems including, but not limited to, traffic congestion, safety, health, and pollution. To tackle them, public administrations have implemented roadside infrastructures such as cameras and sensors to collect data about environmental and traffic conditions. In the case of traffic sensor data not only the real-time data are essential, but also historical values need to be preserved and published. When real-time and historical data of smart cities become available, everyone can join an evidence-based debate on the city’s future evolution. The TRAFAIR (Understanding Traffic Flows to Improve Air Quality) project seeks to understand how traffic affects urban air quality. The project develops a platform to provide real-time and predicted values on air quality in several cities in Europe, encompassing tasks such as the deployment of low-cost air quality sensors, data collection and integration, modeling and prediction, the publication of open data, and the development of applications for end-users and public administrations. This paper explicitly focuses on the modeling and semantic annotation of traffic data. We present the tools and techniques used in the project and validate our strategies for data modeling and its semantic enrichment over two cities: Modena (Italy) and Zaragoza (Spain). An experimental evaluation shows that our approach to publish Linked Data is effective.