| [7] |
Mobility to different points of interest (POI) such as restaurants, grocery stores, religious establishments, fitness centers, and supermarkets from SafeGraph, social vulnerability index (i.e., socioeconomic status, household composition and disability status, minority status and language proficiency, housing type and transportation) from CDC, unemployment rate from American Community Survey, population density from WorldPop, and  concentration from NOAA, Daily COVID-19 cases from LA Public Health Department |
Gradient Boosting Decision Tree (GBDT) |
| [45] |
Mobility patterns at retail and recreation, grocery stores and pharmacies, parks, transit stations, workplaces, residential areas, and schools from Google Mobility Report and UNESCO |
Random Forest (RF), K-Nearest Neighbor (KNN) |
| [46] |
Subway ridership and vehicular traffic data from the Metropolitan Transportation Authority (MTA), transit, driving, and walking trips from Apple Mobility Trends reports |
Agent-based simulation (MATSim-NYC), deep-learning-based real-time video processing method (e.g., RetinaNet, ResNet-50) |
| [47] |
Time-lapse webcam images from different urban zones (e.g., tourist areas, residential areas) for detecting pedestrian and their activities |
YOLO, Single Shot MultiBox Detector (SSD), and Faster R-CNN |
| [48] |
Information on policy instruments from the press release from states and mobility data from google mobility reports |
Extreme Gradient Boosting Decision Tree (XGBDT) |
| [49] |
Coronavirus case count from WHO, CDC, NY Times, and Texas DSHS, socioeconomic data from US census, data on mobility changes in retail and recreation, grocery and pharmacy, parks, transit stations, workplaces, residential from Google and Apple |
SIRNET - a hybrid model comprising of Recurrent Neural network (RNN) and SEIR model |
| [50] |
COVID-19 cases from the county, socioeconomic data (e.g., age, gender, education, income) from various ministries, federal states, and municipal governments, built environment (e.g., density of airports, train stations, grocery stores, parks) from OpenStreetMap |
Bayesian Additive Regression Trees (BART) model |
| [51] |
Monthly Chinese tourist arrival data from the 2003 SARS outbreak up to October 2019 from the National Travel & Tourism Office and Australian Bureau of Statistics |
Long Short-Term Memory (LSTM) model |
| [52] |
COVID-19 cases and deaths, number of trips, social distancing index, out of county trips, transit mode share, population density, population staying at home, working from home, socioeconomic information (e.g., race, income, and employment status, unemployment rate) from University of Maryland (UMD), age category, education, male to female ratio, gross domestic product from US Census Bureau |
Ridge, LASSO, Elastic Net modeling techniques |
| [53] |
COVID-19 data from the European Centre for Disease Prevention, mobility in retail and recreation, grocery and pharmacy, parks, transit stations, workplaces, and residential areas from Google and driving, walking, and transit trips from Apple |
Gradient Boosting Decision Tree (GBDT) |
| [54] |
Coronavirus daily incidence from Worldometer, number of searches for Antiseptic, Hand washing, Hand sanitizer, and Ethanol from Google search trend |
LSTM |
| [55] |
COVID-19 cases and deaths from the CDC, POI visits data from SafeGraph, population and income from US census, policies such as the closure of public venues and schools from local government |
COVID-GAN (Conditional generative neural network approach) |
| [56] |
An indoor video data recorded at the Institute of Management Sciences, Hayatabad, Peshawar Pakistan, which contains video sequences captured from the overhead view |
You Only Look Once (YOLO) |
| [57] |
COVID-19 cases and deaths from New York Times, mobility index from Google mobility reports, socioeconomic data (e.g., population, age, number of hospitals and ICU beds, percentage of smokers and diabetes, and heart disease mortality) from CDC |
Hawkes process using the expectation-maximization (EM) algorithm |
| [58] |
Transportation data (e.g., traffic volume, accident) from 73 signalized intersections, COVID-19 cases from Michigan’s official Coronavirus dashboard, weather factors (e.g., temperature and wind speed) from National Oceanic and Atmospheric Administration, social distancing data from UMD, Crash Data from Michigan State Police (MSP) |
Long Short Term Memory (LSTM) |
| [59] |
COVID-19 cases from local COVID-19 Dashboards, census-tract geographic boundaries and socioeconomic attributes from US Census Bureau, POIs with human travel patterns from SafeGraph |
Walktrap network-based community detection method, modified SEIR model, Ensemble Kalman Filter |
| [60] |
Coronavirus cases, deaths, patient age, gender from the World Health Organization and John Hopkins University |
Decision Tree (DT) Classifier, Support Vector Machine (SVM) Classifier, Gaussian Naïve Bayes Classifier, Boosted RF Classifier |
| [61] |
Coronavirus cases and stay-at-home order from the New York Times, socio-economics (e.g., population density, labor force rate, unemployment rate, income) from US Department of Agriculture, Economic Research Service, weather information (e.g., temperature, wind speed, precipitation, solar radiation) from gridMET, mobility data from Google mobility reports |
Elastic net (EN) model, KNN, DT, RF, GBDT, and Artificial Neural Network (ANN) |
| [62] |
Mobile device location data based on GPS, detailed call record, Cellular Network data, Social media Location-based Services, COVID-19 data from Johns Hopkins University, socioeconomic data from American Community Survey, travel information from National Household Travel Survey |
Deep learning algorithms |
| [63] |
Migration data (e.g., migration index, % of migrants coming from other cities, and internal travel flow index) from the Baidu Maps LBS open platform, confirmed cases |
Ensemble-based back propagation neural network (BPNN) model |
| [64] |
Car traffic data from the road management center of Iran, air flight traffic data from Iranian’s airports and air navigation office, and COVID-19 daily new cases from Johns Hopkins University |
LASSO regression, KNN, RF, and ensemble learning |
