Search

This study presents a comparative analysis of three predictive models with an increasing degree of flexibility: hidden dynamic geostatistical models (HDGM), generalised additive mixed models (GAMM), and the random forest spatiotemporal kriging models (RFSTK). These models are evaluated for their effectiveness in predicting PM 2.5 concentrations in Lombardy (North Italy) from 2016 to 2020. Despite differing methodologies, all models demonstrate proficient capture of spatiotemporal patterns within air pollution data with similar out-of-sample performance. Furthermore, the study delves into station-specific analyses, revealing variable model performance contingent on localised conditions. Model interpretation, facilitated by parametric coefficient analysis and partial dependence plots, unveils consistent associations between predictor variables and PM 2.5 concentrations. Despite nuanced variations in modelling spatiotemporal correlations, all models effectively accounted for the underlying dependence. In summary, this study underscores the efficacy of conventional techniques in modelling correlated spatiotemporal data, concurrently highlighting the complementary potential of Machine Learning and classical statistical approaches.

We propose a twofold adjustment for Event Studies considering spatiotemporal data in a multivariate time series framework where the data are characterized by spatial and temporal dependence. The first adjustment consists of modeling the spatiotemporal dynamics of the data by implementing several geostatistical models capable of handling both spatial and temporal components, as well as estimating the relationship between the response variable and a set of exogenous factors. With the second adjustment, we propose to use cross-sectional-adjusted test statistics directly accounting for spatial cross-correlation. The proposed methods are applied to the case of NO 2 concentrations observed in Northern Italy during the first wave of the COVID-19 pandemic. The key findings are as follows. First, all the considered geostatistical models estimate larger reductions in the major metropolitan and congested areas, while smaller reductions are estimated in rural plains and in the mountains. Second, the models are nearly equivalent in terms of fitting and are capable of identifying the true event window. Third, by using spatiotemporal models we ensure the residuals are uncorrelated across space and time, thus allowing Event Studies test statistics to provide reliable and realistic estimates. Fourth, as expected, all test statistics show significant reductions in NO 2 concentrations starting from the first few days of lockdown. Supplementary materials accompanying this paper appear online.

Unfortunately, original online version of this article's, one of the references was wrongly fixed and it has been fixed with this corrections as follows: ‘Fassò, A., Rodeschini, J., Fusta Moro, A., Shaboviq, Q., Maranzano, P., Cameletti, M., Finazzi, F., Golini, N., Otto, P. (2023) Agrimonia: a dataset on livestock, meteorology and air quality in the Lombardy region, Italy. Sci Data 10, 143 (2023). The original article has been corrected.

In this paper, we employ a qualitative and quantitative analysis of European Union (EU) economic policies that were implemented to navigate the shocks generated by the 2020 COVID-19 pandemic and the 2021-2022 inflationary crisis. We conduct a country-by-country analysis of Italy, France, Germany, Spain, and the Eurozone to investigate the macroeconomic and structural effects of the pandemic and inflation. Comparing the financial and economic crises of the 2007-2009 period and the two crises of 2020-2022, we demonstrate that EU institutions have altered approaches toward managing socioeconomic shocks and crises. Specifically, we observe rapid and collaborative fiscal policy responses that kept investment and labor markets resilient. In particular, sustaining investment investments through public spending has made it possible to contain the losses generated by the pandemic. Moreover, we show that the inflationary crisis began well before the Russia-Ukraine war and that its origins are related to the techno-economic transition induced by the commodities race and the excessively long international value chains. However, the absence of a common European public economic policy based on a common public budget may exacerbate the structural challenges that the EU must address; in particular, the green and digital transition and global geopolitical rearrangements. The sustainability of the EU is linked to the choice of whether to become a lead global player with a cohesive economic and fiscal structure or to privilege internal individualism and become increasingly less incisive.

We propose a novel model selection algorithm based on a penalized maximum likelihood estimator (PMLE) for functional hidden dynamic geostatistical models (f-HDGM). These models employ a classic mixed-effect regression structure with embedded spatiotemporal dynamics to model georeferenced data observed in a functional domain. Thus, the regression coefficients are functions. The algorithm simultaneously selects the relevant spline basis functions and regressors that are used to model the fixed effects. In this way, it automatically shrinks to zero irrelevant parts of the functional coefficients or the entire function for an irrelevant regressor. The algorithm is based on an adaptive LASSO penalty function, with weights obtained by the unpenalised f-HDGM maximum likelihood estimators. The computational burden of maximisation is drastically reduced by a local quadratic approximation of the log-likelihood. A Monte Carlo simulation study provides insight in prediction ability and parameter estimate precision, considering increasing spatiotemporal dependence and cross-correlations among predictors. Further, the algorithm behaviour is investigated when modelling air quality functional data with several weather and land cover covariates. Within this application, we also explore some scalability properties of our algorithm. Both simulations and empirical results show that the prediction ability of the penalised estimates are equivalent to those provided by the maximum likelihood estimates. However, adopting the so-called one-standard-error rule, we obtain estimates closer to the real ones, as well as simpler and more interpretable models.

When time series are observed with noise, the popular Augmented Dickey–Fuller (ADF) unit root test and Johansen's cointegration test are oversized: the ADF tends to conclude for stationarity too often and Johansen's test finds too many cointegrating relations. This fact is well-known but no simple solution has been proposed in the literature. In this work, we show why this happens and prove theoretically and by Monte Carlo simulations how three different filtering approaches can significantly improve the performance of the two tests applied to noisy data without harming their properties when observations are free from noise. We show how conclusions can change when using filtered time series in two real applications: one concerning wholesale electricity prices in European countries, and the second warning against pairs trading strategies based on spurious cointegrating relations among stock prices.

The air in the Lombardy region, Italy, is one of the most polluted in Europe because of limited air circulation and high emission levels. There is a large scientific consensus that the agricultural sector has a significant impact on air quality. To support studies quantifying the role of the agricultural and livestock sectors on the Lombardy air quality, this paper presents a harmonised dataset containing daily values of air quality, weather, emissions, livestock, and land and soil use in the years 2016–2021, for the Lombardy region. The daily scale is obtained by averaging hourly data and interpolating other variables. In fact, the pollutant data come from the European Environmental Agency and the Lombardy Regional Environment Protection Agency, weather and emissions data from the European Copernicus programme, livestock data from the Italian zootechnical registry, and land and soil use data from the CORINE Land Cover project. The resulting dataset is designed to be used as is by those using air quality data for research.

Many countries, including Italy, have experienced significant social and spatial inequalities in mortality during the Covid-19 pandemic. This study applies a multiple exposures framework to investigate how joint place-based factors influence spatial inequalities of excess mortality during the first year of the Covid −19 pandemic in the Lombardy region of Italy. For the Lombardy region, we integrated municipality-level data on all-cause mortality between 2015 and 2020 with 13 spatial covariates, including 5-year average concentrations of six air pollutants, the average temperature in 2020, and multiple socio-demographic factors, and health facilities per capita. Using the clustering algorithm Bayesian profile regression, we fit spatial covariates jointly to identify clusters of municipalities with similar exposure profiles and estimated associations between clusters and excess mortality in 2020. Cluster analysis resulted in 13 clusters. Controlling for spatial autocorrelation of excess mortality and health-protective agency, two clusters had significantly elevated excess mortality than the rest of Lombardy. Municipalities in these highest-risk clusters are in Bergamo, Brescia, and Cremona provinces. The highest risk cluster (C11) had the highest long-term particulate matter air pollution levels (PM2.5 and PM10) and significantly elevated NO2 and CO air pollutants, temperature, proportion ≤18 years, and male-to-female ratio. This cluster is significantly lower for income and ≥65 years. The other high-risk cluster, Cluster 10 (C10), is elevated significantly for ozone but significantly lower for other air pollutants. Covariates with elevated levels for C10 include proportion 65 years or older and a male-to-female ratio. Cluster 10 is significantly lower for income, temperature, per capita health facilities, ≤18 years, and population density. Our results suggest that joint built, natural, and socio-demographic factors influenced spatial inequalities of excess mortality

Il presente manoscritto cerca di contribuire alla filosofia della Citizen science, basata sulla collaborazione tra istituzioni pubbliche, ricerca e cittadinanza attiva, offrendo strumenti statistici avanzati per il monitoraggio della qualità dell'aria e per l'analisi delle azioni di protezione ambientale. Il focus principale sarà sulla modellazione temporale, che prevede quindi l'utilizzo di metodologie statistiche per l'analisi delle serie temporali, della qualità dell'aria in Lombardia, cuore economico-finanziario del Nord Italia, che da decenni soffre di un elevato inquinamento e gode di una scarsa qualità dell'aria nonostante gli sforzi compiuti dalle istituzioni. In questo contesto introdurremo e discuteremo nuovi strumenti con l'obiettivo di stimare l'impatto sulla qualità dell'aria che respiriamo di eventi straordinari e shock inaspettati. In particolare, prenderemo in considerazione sia eventi artificiali, come l'introduzione di nuove zone a traffico limitato nelle aree urbane, sia naturali, come la chiusura forzata di tutte le attività produttive e le restrizioni alla mobilità individuale a seguito degli eventi della pandemia COVID-19 nel 2020. Nel complesso, il lavoro ha diversi obiettivi: da un lato, intende offrire una valutazione tecnico-scientifica dei suddetti eventi al fine di convalidare le scelte di politica ambientale della regione e di aprire un dibattito sulle politiche pubbliche basate su evidenze scientifiche e su approcci datadriven; dall'altro, cerca di importare da altre realtà, come il mondo finanziario, tecniche statistiche utili all'analisi dei dati sulla qualità dell'aria, aprendo nuovi fronti di ricerca; infine, vuole mettere a disposizione degli utenti internazionali su larga scala i dati sulle condizioni ambientali della regione Lombardia attraverso lo sviluppo di un software specifico per il download, la visualizzazione e la gestione di queste informazioni. Alcune delle nuove proposte metodologiche discusse nel documento so, The present manuscript attempts to contribute to the Citizen science philosophy, based on the collaboration between public institutions, research and active citizenship, by offering advanced statistical tools for air quality monitoring and for the analysis of environmental protection actions. The main focus will be on the temporal modelling, hence involving the use of statistical methodologies for time series analysis, of air quality in Lombardy, the economic-financial heart of northern Italy, which has suffered from high pollution for decades and enjoys poor air quality despite the efforts made by the institutions. In this context we will introduce and discuss new tools having the aim of estimate the impact on the air quality we breath of extraordinary events and unexpected shocks. In particular, we will consider both artificial events, such as the introduction of new restricted traffic zones in urban areas, and natural, such as the forced closure of all productive activities and restrictions on individual mobility following the events of the COVID-19 pandemic in 2020. Overall, the work has various objectives: on the one hand, it intends to offer a technicalscientific evaluation of the above events in order to validate environmental policy choices in the region and to open a discussion on public policies based on scientific evidence and datadriven approaches; on the other hand, it attempts to import from other realities, such as the financial world, statistical techniques useful for the analysis of data on air quality, opening up new research fronts; eventually, it wants to make data on environmental conditions in Lombardy region available to international users on a large scale through the development of a specific software for downloading, visualising and managing this information. Some of the new methodological proposals discussed in the paper are: the development of an algorithm, presented as a pipeline chain of distinct statistical methodologies, covering all