CovidClue™ Beta test release 0.4
Covid-19 statistic definitions are:
  • Cumulative Deaths are “Deaths - Aggregated death toll” as reported by JHU CSSE and "Deaths" as reported by the New York City Health Department.
  • Cumulative Cases are “Confirmed - Aggregated case count” as reported by JHU CSSE and "Confirmed Cases" as reported by the New York City Health Department.
  • Cumulative Cases Per Hundred Population are (Cumulative Cases)/Population * 100
  • Cumulative Deaths Per Hundred Population are (Cumulative Deaths)/Population * 100
  • Cumulative Case Fatality Rate is Cumulative Deaths/ Cumulative Cases
  • New Cases are the difference in Cumulative Cases in subsequently reported days
  • New Deaths are the difference in Cumulative Deaths in subsequently reported days
  • New Cases-7 Day Average are the seven-day rolling average of New Cases
  • New Deaths-7 Day Average are the seven-day rolling average of New Deaths
  • New Cases per Million are (New Cases-7 Day Average/Population) * 1,000,000
  • New Deaths per Million are (New Deaths-7 Day Average/Population) * 1,000,000
  • Estimated Infections by Deaths, IFR is calculated by dividing Cumulative Deaths by an Infection Fatality Rate of 0.006. (0.6%).
  • Estimated Percent of Population Infected by Deaths, IFR is calculated by dividing Estimated Infections by Deaths, IFR by area population
    1. For CovidClue™ we use Azure Maps, a cloud-based service that automatically assigns color gradients to the COVID-19 statistics we upload. When a small area of the U.S. experiences Cases or Deaths far in excess of the national average, it can cause color gradients that are difficult to interpret. As of December 31, 2020, we switched to a new methodology that takes the square root of the COVID statistics before color gradients are assigned, resulting in displays that users should find more intuitive and meaningful.
    2. COVID-19 data for the five boroughs of New York City (Manhattan, Bronx, Brooklyn, Queens, and Staten Island) is lagged by 3 days plus an additional day for reporting. This causes the most recent daily data for New York City to read zero and the 7-day rolling averages for cases and deaths to read incorrectly when the “End Date” is set within the past four days. To work around this data reporting issue, reset the “End Date” in the user interface back by five days prior to the current date when viewing data that includes New York City.

    Statistics in CovidClue™ are derived from Covid-19 observations complied in the JHU CSSE COVID-19 Dataset and New York City Health Department "boroughs-case-hosp-death.csv" data file.

    Each night the Center for Systems Science and Engineering (CSSE) uploads compiled observations on confirmed COVID-19 cases and deaths to GitHub in Comma Separated Value (CSV) format. At 4am ET we download the CSSE data. At 4:05am we download the New York City Coronavirus Disease 2019 (COVID-19) data. Our database update is usually completed by 4:30am.

    COVID-19 data is organized by County, State, and Country. We validate COVID-19 data against standard geographic definitions provided by the U.S. Census Bureau. Once this validation is accomplished, the data can be further organized by Core-Based Statistical Areas (CBSA or “Metro Areas”) and compared to population estimates provided by the Census Bureau. The delineations for the CBSA are also supplied by the Census Bureau.

    A key feature of CovidClue™ is analysis by metro areas, or Core-Based Statistical Areas (CBSA), as delineated by the U.S. Office of Management and Budget (OMB). A CBSA is an urban area consisting of one or more counties that are tied together by commuting and other economic activity. As such, CBSA can be better geographies for analysis of the COVID-19 pandemic than states or counties, which have political boundaries. More about the delineation of CBSA can be found in this OMB bulletin, 2010 Standards for Delineating Metropolitan and Micropolitan Statistical Areas. In order to analyze COVID-19 by CBSA, the data must be validated and coded by county Federal Information Processing Standards (FIPS) codes.

    Our aggregated statistics for areas outside of New York City generally tie to the statistics published by the JHU Coronavirus Research Center. The states of Georgia, Tennessee, Michigan, and Hawaii report COVID-19 statistics for "Out of State" cases; we do not include the "Out of State" cases in CovidClue because these cases are not coded by FIPS county code. As a result, our aggregations for these states are slightly lower than those published by JHU.

    Our aggregated statistics for New York City (five boroughs of Manhattan (New York County), Brooklyn, Bronx, Queens, and Staten Island (Richmond County)) tie to the "data-by-day.csv" data file as published by the New York City Health Department. This data is published with a three-day lag, due to the standard delays in reporting for the New York City Health Department. This will have an effect on our 7-day averages. These aggregated statistics are slightly lower than the citywide Cases, Hospitalizations, and Deaths published on the Health Department website because of records with missing geographic information.

    Population statistics are sourced from "Population Estimates and Projections" (2019) by the U.S. Census Bureau and are for the year 2019. We use the same racial and ethnic categories as reported in the data for "race alone." Please note that "Hispanic" is an ethnicity, not a race, so the racial categories plus "Hispanic" do not total to our "Minority" category. The Census Bureau does not publish statistics for "Minority" population; we calculate this category by subtracting "Non-Hispanic White Alone Male" plus "Non-Hispanic White Alone Female" from the total population. This approach is consistent with 13 CFR § 124.103.

         Minority = TOT_POP -  (NHWA_MALE + NHWA_FEMALE)

    Commuting, health insurance, and income statistics are from the 2018 American Community Survey (ACS) by the U.S. Census Bureau. The ACS is a yearly survey that covers only the larger metro areas (CBSA) and counties. As a result, the reports will show "0", "9999", or "99%" when ACS data is not available. The commuting percentages are calculated with the subcategory in the numerator and the survey total in the denominator, less those who "Work at Home." Commuting patterns have changed since the ACS survey was taken in 2018 but we believe the data is still indicative of the need for essential workers to use public transit.