Epimodel data repository for COVID-19

This is the repository for sharing data in the Epidemics COVID-19 project.

While the data can be useful as is, the library and toolkit to use it is epimodel, this repo is updated asynchronously to it.

All data files (mostly CSV and HDF5) are indexed by region Code (ISOalpha2 code for countries, ISO 3166-2 for country subdivisions, see the main repo readme for details).

NOTE 2020-04-6: There has been renaming of the files to be more understandable, apologies for any inconvenience.

Issue tracking

Please use the covid repository to submit issues.

Common data

• regions.csv - Common region set used by epimodel
• regions-gleam.csv - Additional regions with the GLEAM basins (only useful for GLEAM) (Skipped: country Norther Cyprus with city Ercan, country Kosovo with city Pristina) (Note: the populations may be missing or wrong, esp. for small regions!)

Updated

• foretold.csv - Infected estimates from Foretold predictions
• johns-hopkins.csv - John hopkins data, recoded from CSSE github

Static (or seldom updated)

• timezones.csv - Countries (by ISO alpha-2 code) and their timezones
• rates.csv - Countries (by ISO alpha-2 code) and the following based on demographic data from the UN and doi:10.1101/2020.03.09.20033357 Hospitalization fraction of COVID-19 infected needing hospitalization Critical fraction of COVID-19 infected needing intensive care CaseFatalityRate fraction of COVID-19 infected that would die (IFR or infection fatality rate in the paper)
• verity_et_al_rates.csv - Various rates from doi:10.1101/2020.03.09.20033357
• age_dist_un.csv - UN Age distribution for various regions, estimated absolute numbers from https://population.un.org/wpp/
• estimated-infectious.csv - The historical estimations done during the testing. (Countries sorted by ISO aplha-2 code)
• hospital-capacity.csv - The capacity of hospital beds counted per 100k citizens with capacity of new infectionus per day per 1000 citizens. Year and source URL attached. (Countries sorted by aplha-2 code)
• max_percentage_of_infected_population_for_enough_icu.csv - The maximal percentage of the infected the infected population for enaugh ICU beds.

EpidemicForecasting.org countermeasure DB

Direct conversion from notion.se CSVs, updated

• countermeasures-features.csv - 24 selected features
• countermeasures-model-0to1.csv - 9 gradual featureg (pre-processed, early selection for modeling)

Not updated

• countermeasures-selected-binary.csv - 16 selected binary features (not updated)
• countermeasures-model-0to1-split.csv - Split of countermeasures-model-0to1.csv into binary flags (for CM strength modeling). All degrees of a feature up to current level are active.
• countermeasures-model-0to1-split_exclusive.csv - Same as above, only the current degree of every multi-level feature is active.

Notes: Other data sources

For reference

• Polymath COVID-19 data wiki page

Unused

• http://www.geopostcodes.com/index.php?pg=resources&file=ISO_3166_2_Regions_codes#iso
• https://gist.github.com/mindplay-dk/4755200 - marked as defunct and full of errors