Table 1 List of datasets.

From: Forecasting influenza activity using machine-learned mobility map

Dataset

Source

Year, Temporal resolution

Spatial resolution

ILI Emergency Department visits in New York City (NYC)

EpiQuery: NYC Syndromic Surveillance

2016–2017 flu season (daily)

County level

ILI Flu-A positive %

CDC FluView

2016–2017 flu season (weekly)

HHS Region 2, State of New York

ILI Lab tested flu positive counts for State of New Jersey (NJ)

The New Jersey Department of Health

2016–2017 flu season (weekly)

County level

Influenza positive counts for Australia (AUS)

National Notifiable Disease Surveillance System, Australia Government, Department of Health

2016 flu season (daily)

State level

Aggregate mobility flows (AMM)

Google

2016–2017 (weekly)

County level (NY, NJ), State level (AUS)

NY, NJ Commuter counts (COMMUTE)

American Community Survey

2009–2013 (typical day)

County level

Interstate commuter flows in Australia

Australian Labor Market Statistics

2006 census

State level

NY, NJ population

U.S. Census Bureau

2013 population estimates

County level

Australia population

Australian Bureau of Statistics

2016 population estimates

State level

  1. Each dataset is provided along with the source, temporal, and spatial resolution. The first four datasets pertain to influenza incidence rate monitoring, while the remaining are used to model movement between counties/states. ILI stands for Influenza-Like Illness, which includes influenza and other illnesses that present similar symptoms. Clinical lab tests are used to confirm whether it is influenza, and if so, to identify the particular strain. During a typical influenza season, multiple strains circulate in the population, and Flu-A positive% is the percentage of lab-tested influenza specimens that tested positive for Influenza A. A full list of references are provided in the Data Availability Statement.
Back to article page