| [70] |
COVID-19 death cases from US Facts, Pollutants data (e.g.,  , benzene, formaldehyde, acetaldehyde, carbon tetrachloride) from Environmental Protection Agency and Centers for Disease Control and Prevention, weather (e.g., temperature, precipitation, sunlight and UV exposure), land cover, health status (e.g., disabled, obese, overweight) from Centers for Disease Control and Prevention, socio-economics (e.g., health insurance, poverty, income) and commuting information (e.g., travel modes, time) from the US Census |
Geographical weighted RF (GW-RF) |
| [71] |
COVID-19 epidemiology data from [128] and New York Times COVID-19, Daily air traffic (people/day) from International Air Transport Association. |
Susceptible-exposed-infectious-recovered (SEIR) models, Bayesian Interference models |
| [83] |
Flight data from Bureau of Transportation Statistics, ground traffic data from NYC Open Data, air pollutant data (e.g., CO,  , Ozone, and  ) from Aura Satellite (OMI instrument) and Environmental Protection Agency |
Support vector machine (SVM) |
| [91] |
 ,  , and  concentrations from the Secretary of the Environment of the Municipality of the Metropolitan District of Quito |
Parametric analysis |
| [96] |
Ground-based  concentration from central pollution control board (CPCB), satellite-derived MODIS Aerosol Optical Depth (AODs) data, and meteorological data (wind speed, temperature, rainfall, relative humidity, and mixing height) from Indian Meteorological Department. |
Artificial neural network (ANN) |
| [97] |
Meteorological data (e.g., temperature, precipitation, wind speed, wind direction, and air pressure), air quality data for the years 2014–2020, and lockdown data from the Austrian government |
Principal components analysis (PCA), random forest (RF) |
| [98] |
Eight meteorological parameters (e.g., surface wind components, surface temperature, relative humidity, cloud coverage, precipitation, pressure, and planetary boundary layer) from government agencies |
Extreme Gradient Boosting Decision Tree (XGBDT) |
| [93] |
Daily minimum and maximum temperature, average wind speed and direction, average relative humidity, daily cumulative precipitation, and  and  concentration from ARPA Lombardia, and time and seasonal variables |
XGBDT |
| [99] |
Meteorological data (e.g., temperature, pressure, wind speed, cloud cover, solar radiation, ultra-violate radiation) from ERA5 reanalysis dataset, and  data from the Earth Sciences Department of the Barcelona Supercomputing Center. |
Gradient Boosting Decision Tree (GBDT) |
| [100] |
CO,  ,  ,  ,  , and  data from six monitoring stations between March and April |
Reduced-spaced Gaussian Process Regression and ANN |
| [101] |
 ,  ,  ,  ,  ,  ,  , Toluene, benzene, and NH3 data from the Central Pollution Control Board and Ministry of Health and Family Welfare (MoHFW) |
Decision tree (DT), RF |
| [102] |
 data from monitoring station of the US Embassy, Dhaka,  ,  , CO, and  from AirNow,  measured by the Copernicus Sentinel-5 Precursor Tropospheric Monitoring Instrument |
Generalized additive models (GAMs), wavelet coherence, RF |
| [103] |
Daily COVID-19 cases and lockdown level from Statistical Portal of São Paulo State, and meteorological variables (e.g., relative humidity, maximum temperature, atmospheric pressure, wind speed, and global solar radiation), CO,  ,  , NO,  , and  from Environmental Company of São Paulo State database (CETESB) |
ANN models (Multilayer Perceptron overview, Radial basis function, Extreme Learning Machines, Echo State Networks) |
| [104] |
The daily average of  and  from environmental monitoring stations located in the cities, COVID-19 death, resuscitations, and hospitalization from the French National Public Health Agency |
ANN |
| [105] |
Italian Civil Protection, Regional Environmental Protection Agencies (ARPA) |
SVM, K-Nearest Neighbor (KNN), GBDT, Classification and regression tree (CART), RF, Multilayer perceptron (MLP), Ada boosting with decision tree (AdaBoost), Extra tree (ET) |
| [106] |
Nine observation sites in Hangzhou, China |
RF |
| [107] |
City-level hourly data of 4 pollutants from Qingyue Open Environmental Data Center, meteorological data (e.g., temperature, relative humidity, wind direction, wind speed, and air pressure) from “worldmet” R package |
RF, New augmented synthetic control method |
| [108] |
COVID-19 positivity, mortality, and total case count from Italian Civil Protection, air pollutants (i.e.,  ,  ,  ,  , CO, Benzene, and  ) from Italian Ministry of Agriculture, Food and Forestry and Regional Environmental Protection Agency (ARPA), air pollution from Italian National Institute of Statistics |
RF |
| [109] |
Emission factors from different sources (open access and near-real-time measured activity data, proxy indicators and other available reports), stringency index from Oxford COVID-19 Government Response Tracker (OxCGRT) |
GBDT |
| [110] |
Pollutant ( ,  , and  ) emission from Open Government Data (OGD) and World Bank, per capita GDP from Federal Reserve Bank of St. Louis |
ML-based complex causality algorithm (D2C) |
| [111] |
Pollutants (e.g.,  , CO, and  ) data were captured by MONICA (a cooperative air quality monitoring station) and transmitted via a Bluetooth serial interface to a Raspberry Pi Mod. 3 + based datasink with Raspbian OS |
Shallow neural network (SNN) |
| [112] |
Pollutants (e.g.,  ,  ,  , and CO) data from China National Environmental Monitoring Center, meteorological data (e.g., wind direction and speed, temperature, relative humidity, and pressure) from NOAA Integrated Surface Database |
RF |
| [113] |
 data from operational Copernicus Sentinel 5 Precursor (S5P) TROPOMI, CAMS regional air quality models, European Centre for Medium-range Weather Forecasts (ECMWF) |
GBDT |
