Data definitions and details

For questions related to this documentation or the website, dashboards, and associated maps, data and applications, please email us at Contact@FloridaCovidAction.com.

For the current data definitions and schema information for DOH-published data, please contact DOH directly at COVID-19@FLHealth.gov. The most recent PDF version of this information can be downloaded here. Data obtained from DOH for the Florida COVID Action website can be found here.

Most data maintained by DOH is updated every day, typically between 11 AM – 12 PM ET (UTC-4). We update any data based on DOH data shortly thereafter, no later than 2 PM ET (UTC-4) each day. If DOH data is delayed, our data will be delayed, as well, since we insist on using the most authoritative data available.

All data feeds provided by Florida COVID Action

Other Florida COVID Action APIs:

All Public DOH service layers:

Please note: DOH has moved, deleted and created new versions of their public data feeds often in the past month. We will try to keep up with them and update the links below as soon as we are able.

External data feeds used on the dashboard

Historical COVID Data for Florida:

How our map is different:

We count everyone who has had a confirmed-positive COVID-19 lab result, including antibody testing. Any cumulative count of positive people in Florida should include any person who has a confirmed-positive lab result, whether tested while symptomatic or not. This helps us better track the disease burden on healthcare systems and its spread in the community.

We show the number of positive people, hospitalizations, and deaths for everyone tested in Florida, regardless of their legal address. DOH excludes non-residents from its demographic data, zip code data, death data, and case data over time.

We apply the CDC’s guidance on counting “recovered” cases to show by-county estimates of the number of recovered people. This calculation includes all cases diagnosed more than six weeks ago where no death, hospitalization or ER admission is on file, for both residents and non-residents.

We show hospital data – including active bed counts, capacity, and ICU/isolation availability, updated hourly by the Agency for Health Care Administration (AHCA) through a public data feed.

We also break down deaths and hospitalizations by age group for both residents and non-residents.

We provide a report card for each county to measure readiness to enter the next phase of reopening based on the criteria DOH and the State outlined in late April.

We provide data about the impacts of COVID-19, including lives lost, economic, and social factors by region and county.

We provide testing site locations and data, including hours of operation, test type, and contact information for every testing facility in the state. We have a separate testing application tool, as well, which can be used to search for facilities and print directions.

We provide information about community resources, including the locations, contact information and hours for Red Cross facilities, food banks, USDA Meals for Kids locations, and more.

We show city-level case data, made available through daily PDF reports but never added to the DOH dashboard, so communities can better-understand their risk.

We calculate positivity based on the number of people tested, not the number of new cases divided by the total numbers of tests processed across all labs each day, which includes duplicates and re-tests (though we also provide this information for context).

We put the data in context, clearly stating each variable’s limitations and caveats, for all the data on the site, including DOH’s data.

We provide public data feeds that aren’t dependent on the now-unreliable DOH dashboard, and are pulled directly from the PDF reports issued by DOH each morning. Anyone can download the raw data and analyze it, regardless of whether or not the dashboard is crashing, and we provide the daily data for all PDFS and data tables published by DOH each day on our Open Data site.

Most importantly, we ask the community for input. If there’s data the public wants to see, and we have a way of accurately counting or finding that data, we will do our best to do so. If people want to add event information or resources the public can access for testing or for help, they can add it to our map by filling out a short, anonymous survey.

What we still can’t do:

Count probable cases and deaths. Florida does not follow CDC guidance on reporting probable cases, hospitalizations or deaths by not publishing this data. Without critical information about probable cases, our understanding of the virus, its burden on our communities and health systems, and its change in spread over time is very limited.

Measure precise recovery and recovery rates based on case data. The CDC outlined clear guidance for determining what qualifies someone as being “recovered” two months ago, but DOH never released that data (even though they have it). Without the data from DOH, we can only estimate the number of recovered based on the data DOH already makes public, which could be over/under the CDC-outlined guidance.

Show total cases that include non-residents at the zip code or city level. Both the zip code and city-level data provided by DOH excludes non-residents and all persons testing positive by antibody testing.

Provide county or zip-code level data for antibody testing. Data currently is submitted based on county, presumably the county of residence, but that is not clear. It is also unclear if this data includes residents and non-residents.  

Show data about cases in prisons, jails and other detention centers. DOH currently only reports prison deaths of confirmed-positive inmates. No data is provided regarding case totals, testing, or probably cases for prisons or jails in the state of Florida.

What I’m looking to find:

  1. Data on community resources – including (but not limited to – food banks, elder care, transportation assistance, assistance with pets and children, etc., from county to local level. If a fire station is giving away boxed lunches, let me know. If a closed school is providing technology, books or food for kids, let me know. If a group of people decided to form a volunteer dog-walking crew to help those with restricted mobility or those advised not to go out (whether positive or vulnerable), let me know.
  2. Local data on deaths, hospitalizations and cases that may not be published by the Florida DOH.
  3. Your stories, how this virus has impacted you and your family, and your hopes going into the future.

The form below is anonymous (optional to enter your email if you’d like me to follow up). Please feel free to post, share, and add data!

Frequently asked questions

Q: Why is your total positive people count so much higher than DOH’s?
DOH publishes total cases, not positive people. DOH stopped automatically reporting positive new persons in late April, and began waiting to add new people to the total case count until a DOH epidemiologist has time to review the file.

Additionally, cases are not currently created for those who receive positive antibody test results, and so DOH excludes them from that total. We show the total number of people who have definitive lab results showing they have or have had COVID-19 regardless of the type of test. This helps us understand how prevalent the virus has been in our communities. You can view DOH’s weekly antibody report here.

Q: Where does the hospital data come from?
Hospital data comes directly from a live feed of the Agency for Health Care Administration’s hospital census data, which is updated every half-hour with total bed availability for Adult ICU, Pediatric ICU, and isolation rooms.

Q: How are you calculating the number of people recovered?
Our “recovery” count is an estimate and not an exact figure. We apply the CDC’s guidance on counting “recovered” cases to show by-county estimates of the number of recovered people. This calculation includes all cases diagnosed more than six weeks ago where no death, hospitalization or ER admission is on file, for both residents and non-residents.

We still cannot measure precise recovery and recovery rates based on case data with additional data from DOH. The CDC outlined clear guidance for determining what qualifies someone as being “recovered” in April, but DOH never released that data (even though they have it). Without the data from DOH, we can only estimate the number of recovered based on the data DOH already makes public, which could be over/under the CDC-outlined guidance figures.

Q: Where does the antibody data come from?1
DOH provides two datasets related to antibody testing on its website, both of which are updated weekly. One includes data obtained from commercial laboratory testing for COVID antibodies (IgM and IgG) and the other provides information about antibody testing conducted by the state.

DOH has defended its use of using antibodies to count positive persons, although they do not present this data on the dashboard itself, and provides the following statement regarding those tests:

 “The Florida Department of Health is conducting qSARS-CoV-2 (virus that causes COVID-19) antibody testing at several State-run COVID-19 drive-through testing sites test for possible exposure to the virus. All data in this report are from State-operated drive-through testing sites performing POC antibody tests.

The CELLEXTM qSARS-CoV-2 rapid IgG/IgM antibody test is approved by the Federal Drug Administration under an Emergency Use Authorization and the test is designed to detect antibodies against the virus that causes COVID-19. A simple blood draw is used to obtain the sample for testing.”

Data details and explanations

Florida Metrics Data and Calculations

The State plan for reopening outlined the criteria for safely entering the next “phase” in this report. Regardless of what phase a county is currently in, a continued decrease in new cases, continued decrease in emergency room visits for both influenza and COVID-like illness, and either a decrease in the percent of positive labs or percent positivity below 10% must be met for a minimum of two weeks before a county can enter the next phase. These criteria are referred to as the “metrics” data by DOH, and are published through a public API that is updated every Sunday morning.

We collect the metrics data from DOH each week and do a simple measurement to see if all of those criteria are met. If new cases did not decrease in both of the last two reporting weeks, then the county does not meet the criteria to reopen. The same is true for the ER visits for influenza-like illness, COVID-like illness and percent positive labs. Because DOH changed how it calculates the percent of positive labs in late April, nearly all counties show a percent positive of labs below 10%, essentially eliminating the requirement that percent positive labs be decreasing.

Service Layer/API link for our analysis feature layer: https://services9.arcgis.com/pptP3j8cINwF8DAR/arcgis/rest/services/Florida_COVID19_Reopening_Criteria_Analysis_by_County/FeatureServer

Service Layer/API link for the DOH metrics data copied weekly from DOH data:

https://services9.arcgis.com/pptP3j8cINwF8DAR/arcgis/rest/services/Florida_COVID_Action_Metrics_data_from_DOH/FeatureServer

Service layer for DOH’s metrics data (may change if DOH restricts access):

Previous API: https://services1.arcgis.com/CY1LXxl9zlJeBuRZ/arcgis/rest/services/Florida_Health_Metrics/FeatureServer

Current API (as of June 14, 2020): https://services1.arcgis.com/CY1LXxl9zlJeBuRZ/arcgis/rest/services/Metrics_All/FeatureServer

City and Zip Code Data

City data is obtained from the daily PDF report issued by the Department of Health. Case and zip code data only include positive residents and do not include testing data.

DOH notice on city data:

The table includes the cities and counties of residence for cases in Florida residents based on the patient’s ZIP Code. Note that city is not always received as part of the initial notification and may be missing while the case is being investigated

Antibody testing data

DOH publishes two reports on antibody test results every Friday in the morning. We pull this data directly from these reports without changing any information and add it our testing data exactly as it appears in the report.  One report counts commercial antibody tests by county, while the other is a list of antibody testing conducted by DOH based on testing site. Both totals are added into the overall totals for antibodies, with the state-run lab data grouped into the “Unknown” county since no county attribution is provided for those tests.

DOH has confirmed that those who have antibody tests, whether commercial or by DOH, do not overlap with their case totals or with other antibody tests.

Long term care facilities (LTCF) cases and deaths data

The data DOH provides for LTCF cases and deaths is not dated or time-stamped. There is no clear definition of how they are defining cases in their report. The facilities do not have locations aside from county. The case data is updated daily and the deaths data is updated weekly. It is not clear from the definition provided about the deaths data if the data being showed is persons who died in or at LTCFs or if they include residents who contracted the virus at an LTCF (as determined through contact tracing) and later died elsewhere.

LTCF case data notice from DOH:

“Data in this report is provisional and subject to change. The information contained in this report reflects the current available information reported by each facility’s staff to the Department via the Agency for Health Care Administration’s Emergency Status System. The data is not cumulative but reflects the information available for current residents and staff with cases as of yesterday’s date.”

LTCF deaths data notice from DOH:

The information contained in this report reflects the current available information for nursing homes and assisted living facilities that have had a death that can be linked to the facility. The information was reported to the Department’s epidemiological data collection system through the Department’s contact tracing efforts for positive COVID-19 cases. This report will be updated on a weekly basis.

Florida Correctional Institute Deaths

Florida does not currently provide data about cases in its state prison system, and only updates the deaths data, which has been accused of being criminally under-reported, once a week. There is no data currently available regarding cases or deaths in county jails or other correctional facilities.  The weekly PDF report issued by DOH for FCI deaths can be found here.

Note from DOH on FCI deaths data:
“The information contained in this report reflects the current available information for correctional institutions that have had a death that can be linked to the facility. The information contained in this report reflects the current available information reported by each institution to the Department via the Florida Department of Corrections. This report will be updated on a weekly basis.”

How to get access to DOH Case Line data when it’s down

DOH has been inconsistent in updating its case-line data recently. You can either mine the PDF (if it’s been updated), or try a secondary API we’ve generated using the below query:

https://services1.arcgis.com/CY1LXxl9zlJeBuRZ/arcgis/rest/services/Florida_COVID19_Case_Line_Data_NEW/FeatureServer/0/query?where=1%3D1&outFields=*&outSR=4326&f=json
Layer
Florida COVID19 Testing Data Including Antibody Testing
FieldDefinition
PUIs_TotalSum of all persons for whom DOH has a test result on file (PUI)
PUI_PositivePUIs receiving at least one positive lab results
PUI_NegativePUIs receiving only negative or inconclusive lab results
PUI_InconcPUIs receiving only inconclusive lab results
PUI_PendingPUIs who have been tested and are waiting on results, or whom DOH has identified for priority BPHL testing
DOHT_totalThe total number of test results received by DOH (not people)
DOHT_negativeThe total number of negative test results (not people)
DOHT_positiveThe total number of people who received at least one positive lab result
CasesAllThe total number of cases DOH currently has
ALLDEATHSThe total number of people who have died, having previously tested positive for COVID, including residents and non-residents
Antibody_PositiveThe number of people who received positive antibody lab results, either from a commercial/private lab or through DOH
Antibody_NegativeThe number of people who received negative antibody lab results, either from a commercial/private lab or through DOH
Antibody_InconclusiveThe number of people who received inconclusive antibody lab results, either from a commercial/private lab or through DOH
Antibody_TestedThe number of people who received lab results, either from a commercial/private lab or through DOH
Positive_AllSourcesThe number of positive people tested (PUI_Positve + Antibody_Positive)
Negative_AllsourcesThe number of negative people tested (PUI_Positve + Antibody_Positive)
TestedPeople_AllsourcesThe number of total people tested from all PUI and Antibody numbers
Tests_AllSourcesThe total number of tests from DOH by DOHT_total and Antibody_Tested
Percent_PopTestPercent of people who tested positive
Percent_PosTestsPercent of people tested over negative tests
Florida COVID Action Metrics data from DOH
Week_StartDateThe first date of the reporting period week
Week_EndDateThe last date of the reporting period week
W_ILIThe number of emergency room visits with influenza-like illness, exclusive to those who listed a Florida address at the time of intake
W_COVIDLThe number of emergency room visits with COVID-like illness, exclusive to those who listed a Florida address at the time of intake
W_NegTestsThe number of negative test results revceived be DOH
W_PosTestsThe number of new cases created by DOH
W_TTotalThe number of negative tests + newly positive people
W_PercPosThe percent positivity based on W_PosTests divided by W_Ttotal
W_PositivityDOH is no longer using this field
W_EventsNumber of new cases by event date, which is the earliest date of the self-reported onset date, diagnosis date, or test date
Florida COVID19 Reopening Criteria Analysis by County
Week.StartThe first date of the reporting period week
Week.EndThe last date of the reporting period week
M_ILIYes/No: Did ER visits for influenza-like illness decrease during both of the previous two complete reporting weeks
M_COVIDYes/No: Did ER visits for COVID-like illness decrease during both of the previous two complete reporting weeks
M_PosTestsYes/No: Did new cases decrease during both of the previous two complete reporting weeks
D_ILIFor last week only: Whether ER visits for ILI decreased, increased or stayed the same
D_COVIDFor last week only: Whether ER visits for COVID decreased, increased or stayed the same
D_PostTestsFor last week only: Whether new cases decreased, increased or stayed the same
P_ILIThe percent change in ILI ER visits between the last date of the previous two reporting periods only
P_COVIDThe percent change in COVID ER visits between the last date of the previous two reporting periods only
P_PostTestsThe percent change in new cases between the last date of the previous two reporting periods only
Text_ILIText for formatting only
Text_COVIDText for formatting only
Text_PostTestsText for formatting only

I am currently working on the full data definitions document at this time and will upload it once it is complete.

%d bloggers like this: