Impact of Thanksgiving in the US

In [1]:
%config InlineBackend.figure_formats = ['svg']
import datetime
import oscovida as ov

ov.display_binder_link("2020-thanksgiving-us.ipynb")

Execute this notebook with Binder

Get data from Johns Hopkins University (JHU)

In [2]:
cases_jhu, deaths_jhu = ov.get_country_data("US")

Show the standard OSCOVIDA plots together with a black line marking the 26 November 2020 (Thanksgiving in the US).

In [3]:
axes, _, _ = ov.overview(country="US", dates="2020-09-01:2020-12-31");
for ax in axes:
    ax.axvline(datetime.datetime(2020, 11, 26), color="black", linewidth=1)

Discussion

The reported daily new infections and deaths appear to reduce a few days before Thanksgiving, and increase in the weeks after. This is well reflected in the reproduction number (green line for reproduction number for infections, and yellow line for reproduction number for deaths).

The qualitatively similar behaviour of daily deaths and daily infections suggests that one reason for the change around Thanksgiving may be the reporting behaviour: possible reasons could include that numbers have not been passed on during that festive period as quickly as normally, or some people have ignored symptoms just before Thanksgiving, and then went to tests/hospitals a few days later than they would at other times of the year.

On the other hand, if we watch the reproduction number for infections (green line) during November, then there is a downward trend from 1.2 to 1.0 just before Thanksgiving visible. It is possible that Thanksgiving has broken that downward trend.

All in all, it seems difficult to draw firm conclusions.

In [5]:
# axes, _, _ = ov.overview(country="US", dates="2020-11-01:2020-12-21");
# for ax in axes:
#    ax.axvline(datetime.datetime(2020, 11, 26), color="black", linewidth=1)