Project: Aggregating Statistical Models and Human Judgment Config

    Summary: Three models, six forecasts, 12 predictions
    Owner: allicodi
    Model Owners: allicodi, mcandrew
    Time Interval Type Month
    Visibility Public
    Description: This project aims to provide public health officials chimeric forecasts- a combination of probabilistic predictions from computational models, subject matter experts, and trained forecasters- of the COVID-19 outbreak.

    A consensus forecast is an equally weighted combination of probabilistic predictions from individuals. We included consensus forecasts from individual forecasters on Metaculus (https://www.metaculus.com/questions/) and from Good Judgment Open (https://www.gjopen.com/). A chimeric forecast is also included: an equally weighted combination of consensus forecasts and of the COVID-19 Forecasthub ensemble.

    Each month individuals contribute forecasts for incident cases, deaths, hospitalizations, and cumulative number of vaccines at the US national level. Additional targets for forecasting depend on public health needs. Individuals can produce forecasts at the beginning of each month until the middle of the month.

    Files are named YYYYMMDD_YYYYMMDD.csv and correspond to the year, month, and day forecasting begins and ends. All forecasting ends at 2PM EST.
    Home: https://github.com/computationalUncertaintyLab/aggStatModelsAndHumanJudgment_PUBL
    Core data: https://dataverse.harvard.edu/dataverse/aggregating_statistical_models_and_human_judgment
    Truth Data: (No truth data)
    Features: Project: Explorer. Forecasts: Summary, Download.

    Models (3)

    Abbreviation Team # Forecasts Oldest Newest Upload time
    Chimeric Computational Uncertainty Laboratory 6 2021-01-18 2021-06-28 2022-01-18 00:37:28 UTC
    GJOCon Computational Uncertainty Laboratory 0
    MetaculusCon Computational Uncertainty Laboratory 0

    Units (1)

    Abbreviation Name
    US United States

    Targets (12)

    Target Group # Targets Type Outcome Variable Step ahead? Example Target
    2021-01-18 cases 1 continuous number of cases “2021-01-18 cases”
    2021-01-18 deaths 1 continuous number of deaths “2021-01-18 deaths”
    2021-02-22 cases 1 continuous number of cases “2021-02-22 cases”
    2021-02-22 deaths 1 continuous number of deaths “2021-02-22 deaths”
    2021-03-22 cases 1 continuous number of cases “2021-03-22 cases”
    2021-03-22 deaths 1 continuous number of deaths “2021-03-22 deaths”
    2021-04-26 cases 1 continuous number of cases “2021-04-26 cases”
    2021-04-26 deaths 1 continuous number of deaths “2021-04-26 deaths”
    2021-05-31 cases 1 continuous number of cases “2021-05-31 cases”
    2021-05-31 deaths 1 continuous number of deaths “2021-05-31 deaths”
    2021-06-28 cases 1 continuous number of cases “2021-06-28 cases”
    2021-06-28 deaths 1 continuous number of deaths “2021-06-28 deaths”

    Time zeros (6)

    Time Zero Data Version Date # Forecasts Starts Season
    2021-01-18 (No data version date) 1
    2021-02-22 (No data version date) 1
    2021-03-22 (No data version date) 1
    2021-04-26 (No data version date) 1
    2021-05-31 (No data version date) 1
    2021-06-28 (No data version date) 1