docs/data.md

In ELMARDI0107/corona_hel: Tracking the Coronavirus

Data

Minor checks of the crawler.

Sources

• Data from John Hopkins GIS dashboard accessed via readr (used to be from Google sheet).
• Data from Weixin using the nCov2019 package thanks to Guangchuang Yu
• Data from DingXiangYuan scraped using rvest.

Cronjob

To set up the cronjob, recreate your config file somewhere in your home directory.

cd /home
mkdir ncov
R -e "coronavirus::create_config()"
vi _coronavirus.yml

After editing the config file we can create the script that will crawl fresh data. There is a convenience create_script function to do so.

R -e "coronavirus::create_script()"

You should test that all works fine by running the crawler once from the terminal.

Rscript script.R

Then set up the cron job.

crontab -e

In that file place the following to crawl every hour. If you want to set another interval consult crontab guru.

0 * * * * cd /home/ncov && Rscript script.R

All set!

Though every version should be backward compatible, it's good practice to re-run a crawl after installing a new version of the package.

Corrections

:warning: Corrected inaccuracies

v0.1.1

• There was an issue were the number of cases counted in the John Hopkins map of china and world map was wrongly filtered and gave numbers higher than actually are.

v0.0.6

• Corrected death rate from deaths/confirmed to deaths/(confirmed + recovered)

v0.0.3:

• Deaths and recovered numbers for DingXiangYuan data was previously swapped, now fixed.
• Number of suspected by city given by DingXiangYuan is wildly inaccurate, has been removed.
• Replaced Map on JHU tab as color scaling was inaccurate due to timeline.

ELMARDI0107/corona_hel documentation built on April 12, 2020, 12:16 a.m.