Impact of the Euro 2020 championship on the spread of COVID-19

Supplementary code for our Impact of the Euro 2020 championship on the spread of COVID-19 publication.

Usage

Clone with

git clone --recurse-submodules [email protected]:Priesemann-Group/covid19_soccer.git

We also advice you to create an conda environment for this project as we fixed the version of some packages which could interfer with your previously installed packages.

conda create -n "covid19_soccer"  python=3.9
conda activate covid19_soccer
pip install -r requirements.txt

The easiest way to start with our projects is to have a look at the getting started notebook. Please also read our publication!

To fully reproduce our results you may want to download the data too. The data is available here. Alternatively you can also run the analysis on your own this may take some time tho. We do not supply the data for the robustness checks, if you are looking for something specifc feel free to create an issue and ask.

Orientation

Lost? Don't fear, we got your back. If you want to find code to reproduce a specific figure or table from our publication check the Table below:

Most of the figures have one distinct notebooks, the table below gives you a way to find them:

Figure	Title	Notebook
1	Quantifying the impact of the Euro 2020 on COVID-19 spread	overview_figure
2	Example cases illustrate that the spread associated with the Euro 2020 can encompass a substantial fraction of the observed cases	overview_figure
3	What variables can predict the extent of the impact of Euro 2020 matches	correlation_cleaner
S3	Overview of the sum of primary and subsequent cases accountable to the Euro 2020	primary_and_subsequent_fraction
S4	Overview of cases in all considered countries apart from the Netherlands	primary_and_subsequent_overview
S5	We found no significant correlation between cases arising from the Euro 2020 and human mobility	correlation_cleaner
S6	We found no significant correlation between cases arising from the Euro 2020 and the stringency of governmental interventions (NPIs)	correlation_cleaner
S7	We found slight trends in the correlations between the impact of Euro 2020 and the base reproduction number and country popularity	correlation_cleaner
S8	Prediction of the impact of Euro 2020 matches without the two most significant countries in the main model (England and Scotland)	correlation_cleaner
S9	Effect of single Euro 2020 matches on the spread of COVID-19 across competing countries. Whiskers denote 68% and 95% CI.	matches_forest
S10	Including in our model the potential local transmission around the stadium where the matches occur does not significantly increase the overall effect.	beta
S11	A temporal offset of 14 days leads to no inferred effect.	offset_14_days
S12	Changing the days of the match by a large offset results in a non-significant Effect	delay_and_large_offsets
S13	Robustness test for the effect of the temporal association between matches and cases by varying the effective delay	delay_and_offsets
S14	Robustness test for the effect of the width of the delay kernel.	delay_width
S15	Robustness test for the effect of the allowed base reproduction number variability.	change_point_intervals
S16	Robustness test for the effect of the fraction of female participation in football related gatherings	omega_gender
S17	Robustness test for the effect the generation interval.	generation_interval
S18	Robustness test for the remaining priors not studied in the previous figures	other_priors
S19	The combination of the case numbers of England and Scotland leads to similar results	overview_figure_Eng_Sct_combined
S20	Our model is able to identify the delay between infection and reporting of it	delay_and_offsets
S21	Relative popularity of the search term “football” in England and Scotland	google_search
S22	Male-female imbalance over time shows the largest deviations during championship.	imbalance_analysis
S23	The inferred noise terms don’t depend strongly on the length of the analyzed Time-period	compare_long_vs_short
S24	The inferred effect size (percentage of football-related primary infections) don’t depend strongly on the length of the analyzed time-period	compare_long_vs_short
S25-S36	Overview of the posterior for selected countries	extended_overview
S37-S47	Chain mixing of selected parameters for selected countries	chain_mixing

Notes:

You need python>=3.9.

Before you can use the code and rerun the analyses you have to:

• init the submodules:

  #Init
  git submodule init
  # Update package manual (inside covid19_inference folder)
  cd covid19_inference
  git pull origin master

• You may want to download or update the raw data (see here)