totalhack/zillion-covid-19

Use Zillion to analyze COVID-19 data

Zillion COVID-19

This is an example repo showing how one can use Zillion to analyze data related to the COVID-19 outbreak.

The zillion warehouse is defined in zillion_covid_19/covid_warehouse.json. It references a remote sqlite database that is pulled from the Glide COVID-19 repo. The warehouse is configured to ignore the download if the sqlite database already exists locally (check in /tmp for glide_covid_19.db), so it is a one-time snapshot unless you adjust the if_exists setting in covid_warehouse.json or periodically remove your local sqlite DB so it gets refreshed on the next warehouse init.

To see the available warehouse metrics, dimensions, and datasources try the following from the zillion_covid_19 directory:

$ python warehouse.py

"""
---- Warehouse
metrics:
  {
      'cases': <Metric name='cases' type='integer'>,
      'cases_cumsum': <Metric name='cases_cumsum' type='integer'>,
      'cumulative_cases': <Metric name='cumulative_cases' type='integer'>,
      ...
"""

To run reports on the command line use the run_report.py helper script. The following example would show cases and deaths by country for a particular date (not cumulative, there are separate metrics for that):

$ python run_report.py -m cases deaths -d country -c '[("date", "=" , "2020-04-02")]'

"""
             Cases  Deaths
Country
Afghanistan     36       2
Albania         18       1
Algeria        139      28
Andorra         38       1
Angola           0       0
...            ...     ...
Venezuela        3       2
Vietnam         15       0
Yemen            0       0
Zambia           3       1
Zimbabwe         1       0
"""

Use --help for more information. The output is simply a printed dataframe, so data may be truncated in the terminal.

Note: this is meant to be a quick example. Zillion is still in its infancy and is subject to rapid changes.