acab19 is a dead-simple data plotig/browsing tool for COVID-19 datasets
distributed by ourworldindata.
The script uses the coronavirus dataset prepared and distributed by ourworldindata 1. A complete overview of the dataset can be found on owid covid-19-data github page. Here is a quite extensive list of the parameters/data recorded in the dataset.
The script is written in Python and uses the pandas and matplotlib libraries for analyzing and plotting the data.
Since this is a script it does not need installation; you need however to have
all of its dependencies available. Best way to do that, is to install
pipreqs and use it to manage dependancies.
That is in the root folder (aka the folder containing the script), run:
pipreqs .; this will create a requirements.txt file. Then run pip
to install any missing modules: pip install -r requirements.txt. You should
now be good to go!
Data columns/records which are not changing in time are plotted as bar graphs,
where the x-axis is a list of selected countries and the y-axis the values of
the selected data-record. These records are: iso_code,continent,location,
tests_units,population,population_density,median_age,aged_65_older,
aged_70_older,gdp_per_capita,extreme_poverty,cardiovasc_death_rate,
diabetes_prevalence,female_smokers,male_smokers,handwashing_facilities,
hospital_beds_per_thousand,life_expectancy and human_development_index.
E.g. to plot male smokers (%) for Greece, Spain and Italy, use:
acab19.py --countries Greece Spain Italy --x-axis=location --y-axis=male_smokers
which will produce:
When plotting a bar graph, the x-axis must be specified (via a command line argument) and set to
--x-axis=location
Data columns/records which are changing in time, are plotted as (line connected) scatter graphs. The x-axis in this case is the date column and the y-axis is the chosesn data record. For each selected country, we plot one line.
E.g. to plot the fully vaccinated in Greece, Spain, Italy and Israel, use:
./acab19.py --countries Greece Spain Italy Israel --y-axis=people_fully_vaccinated
which will produce something like:
When plotting a scatter graph, the x-axis can be ommited; if not, it must be set to
--x-axis=date
When triggered, the script will download the latest dataset file from ourworldindata
in csv format, aka owid-covid-data.csv.
Users can skip the download (e.g. they have already performed a program run) via the
--csv-file=my_local_owid-covid-data.csv switch.
The script has a couple of switches users can turn on to briefly browse elementary info:
- to see the full list of records/columns use the
--print-columnsswitch, and - to see the full list of countries/locations available, use
--print-countries
Special locations are considered the locations within the dataset that are marked
with an ISO code starting with 'OWID_' (e.g. OWID_AFR). These may be listed
via the --print-countries switch.
The script supports simple queries to search for countries for which static values
meet certain criterea. For example, you may want to plot some (y-)value only for
countries for which the population is within a certain range, or even more complicated
conditions, e.g. countries for which the population and/or the human_development_index
are within a certain range. To do that, you must enter the query string using the
--filter switch. E.g.
-
Only plot countries for which the population is within (8000000, 12000000),
--filter='population>8000000 and population<12000000' -
Only plot countries for which the population is within (8000000, 12000000) and the human development index is larger than 0.8,
--filter='(population>8000000 and population<12000000) and human_development_index>0.8
In general the query string must follow the convention: [some static data value] operator value, where
the static data value can be any column of the static, operator can be aby of >, <, =, >=, <= and
the value is some user defined value (e.g. population>8000000).
You can join miltiple critera using either the or or and operators (aka using an inner/outer join)
and parenthesis (if needed). Examples:
--filter='population>8000000 and population<12000000'
--filter='(population>8000000 and population<12000000) and human_development_index>0.8'
--filter='(population>8000000 and population<12000000) and (human_development_index>0.8 and human_development_index<0.9)'
If you use the filter switch, note that the countries/continents specified via the --countries and
--continents switches will be filtered. If you want all the available countries to be queried, then
leave the --continents and --countries switches unused.
E.g. suppose we want to plot the people_vaccinated_per_hundred column for countries in Asia, Africa or
any of Greece, Italy, Spain, Israel, Belgium, Hungary, for which the population is within
the range (8000000, 12000000) and the human development index is within (0.7, 0.9), then:
acab19.py --filter='(population>8000000 and population<12000000) and (human_development_index>0.7 and human_development_index<0.9)' --countries Greece Italy Spain Israel Belgium Hungary --continent Asia Africa --y-axis=people_vaccinated_per_hundred will produce:
and on screen (STDOUT) you will see that the countries meating the given criterea are Azerbaijan, Greece, Hungary, Jordan, Tunisia and United Arab Emirates.
[1] Max Roser, Hannah Ritchie, Esteban Ortiz-Ospina and Joe Hasell (2020) - "Coronavirus Pandemic (COVID-19)". Published online at OurWorldInData.org. Retrieved from: 'https://ourworldindata.org/coronavirus' [Online Resource]