Analysis of evolution COVID in Belgium

I provide a set of scripts that download the data from Sciensano, the Belgian source for numbers on covid, and subsequently provide basic summary statistics.

WORK IN PROGRESS

Currently you can use the script downloadData.R to download the data from Sciensano and add totals for Belgium and the regions. The other functionality will be added when time allows.

Goal and a warning

This is not a solid statistical analysis, but allows you to browse the data in a bit more detail. The code is free to use (under the GPL-3 license), as long as you refer to me as a source. It comes without any warranty, and should not be used as official communication or proof of anything.

There's enough armchair epidemiologists on social media as it is. Use this code to give you more insight, not to confirm your own firm opinion on how this epidemic evolves. Because frankly, the public data just isn't enough to give us solid answers. It can only serve as a warning.

Organisation

This code base can be run as a shiny app or separate scripts. It's built up as follows :

• scripts: contains the scripts that take the different steps for the summarisation of the data.
• modules: contains shiny modules that are used in the app itself.
• functions: contains a set of utility functions I use throughout the code base.
• Data: contains the data on the population by age class, obtained from https://bestat.statbel.fgov.be/bestat/
• global.R: is the file that sets up the global environment and takes the necessary steps to run whatever script necessary.
• ui.R and server.R form the actual application.

When everything is finished you will be able to run this from within RStudio using :

runGitHub("JoFAM/covidBE_analysis")

Actions

Scripts

• checkPackages.R: checks whether the necessary packages are installed for the shiny app, and loads them if they are.
• create_cumulative.R: creates a dataset for municipalities over the dates by downloading and combining the historical data obtained from Sciensano. This is not part of the app, as it takes quite a while to download. The result is datestamped and stored in the folder Data. Running the script again only downloads the data that is missing in the file.
• createMapPolygons.R: creates a set of sf geometry objects containing the shapes for the administrative regions, and stores them as RDS files. These files are added to the repo, the script is only there for reproducibility.
• downloadData.R: downloads the latest datafiles from Sciensano and prepares the data for the app.

The shiny application takes the following steps:

• download the data from https://epistat.wiv-isp.be/Covid/
• calculate the 7-day running averages to smooth out the effect of differences between week and weekend in reporting. (This is a naive smoothing technique, but gives a better idea of the long term trends)
• save this data with the date as a timestamp. When you rerun the code the next day, the timestamp will tell the code to download the newest data.
• summarise the data in a set of graphs I myself find informative. But you have the data and the scripts, so have a blast.

A word of thanks

• Sciensano for providing the data used in this analysis.
• StatBel for providing shapedata to create maps
• @DaafDP for the ideas about the maps and some interesting code (see https://github.com/DaafDP/Corona)
• @DaafDP also for sending me the data on number of inhabitants and a bunch of shapefiles I used for testing before I found a more reproducible way to add them to this project.