Germany: massively wrong C* values since at least Dec 2021

[fakten-erklaeren]

Irritated by ourworldindata.org showing Germany to have the highest stringency index when other countries have much stricter measures, I checked the data and found that at least half of the C* values are totally wrong (too high):

C2: workplace closing

• OxCGRT value: 3 (require closing (or work from home) for all-but-essential workplaces (eg grocery stores, doctors))
• Real value: 2 (require closing (or work from home) for some sectors or categories of workers)

In reality, there is a work-from-home order only for office desk style jobs ("bei Büroarbeiten oder vergleichbaren Tätigkeiten" - source). So even non-essential workplaces (e.g. manifacturing) are open in general, only some categories of workers have a home office mandate.

Strangely the comment in the data when the value was increased is "Closing of all shops; except essential for those that are not vaccinated; 2G rule applying to all places except for supermarkets and other essential stores". Isn't this off-topic, when this category is about workplaces? Even if all shops would be closed for everyone, this would only be one category of workplaces and according to C2 definition there are still many other categories of workplaces open.

C5: public transport

• OxCGRT value: 2 (require closing (or prohibit most citizens from using it))
• Real value: 0 or 1 (recommend closing (or significantly reduce volume/route/means of transport available))

Actually, there is no recommended closing nor a significantly reduced volume (maybe through distancing requirements, so I'm not sure whether the actual level is 0 or 1), but there is a mask mandate and the "3G rule" (source). This means most people are ALLOWED to use public transport: 73% are fully vaccinated and fulfill the requirements automatically. So value 2, which means the opposite ("most are prohibited"), is clearly wrong.
Again, 3G rule is named as reason for value 2, but the definition of value 2 is clearly NOT met.

C7: restrictions on internal movement between cities/regions

• OxCGRT value: 2 (internal movement restrictions in place)
• Real value: 0 (no measures)

There is ZERO recommendation or even restriction regarding travelling between regions. E.g. the central official list of restrictions does not even mention this topic.
In the dataset, the level was raised to 2 with the comment that the 3G rule was implemented in trains. But this is a matter of public transport measures and has already been counted there. It is a matter of infection risk in crowded transportation and is not at all related to the topic of inhabitants from different regions mixing. Travelling to other cities by e.g. car is not restricted at all and there is no recommendation against it.

C1: school closing

• OxCGRT value: 2 (require closing (only some levels or categories, eg just high school, or just public schools))
• Real value: 1 (recommend closing or all schools open with alterations resulting in significant differences compared to non-Covid-19 operations)

In reality all school categories are generally open, but there are significant differences to normal operation, like mask mandates and requiring vaccination or negative tests. This is actually the comment in the data for setting value 2 ("Berlin introduced a 3G rule for universities; requiring students to have a negative test"), but this is not a closure, or is it?

Others

For C6 (stay at home), the quoted source does not seem to support the value "recommendation to stay at home is still in place". But as this category is pretty soft and could be supported by any statement by politicians done on TV, it might be correct, so I don't intend to dispute it.

Summary

All in all, the average of the C values is currently 82.3, but with the corrections from above, the value is only 55.2!

Could you please correct the wrong data or let me know where I can do so, e.g. by creating a pull request?
Thanks!

Appendix

The list strangely suggesting Germany has strictest Covid measures when actual restrictions are much stricter in other countries:

I'm not sure which values in addition to the C* values are used, but the C* values apparently have the most weight.