Same as always, the weekends get added to the next Monday's results.

This causes spikes, however we can see that these spikes are going down too fortunately.

Let's remove these spike days.

Doesn't really change. Looks like they test more on the beginning of the week and by the end, it drops.

with data_21 as (
select * from (
select country_region, date as date_21, confirmed as confirmed_21, deaths as deaths_21
,LAG(confirmed, 1) OVER (PARTITION BY country_region ORDER BY date ASC) AS prev_day_21
,confirmed - LAG(confirmed, 1) OVER (PARTITION BY country_region ORDER BY date ASC) as new_cases_21
from `bigquery-public-data.covid19_jhu_csse.summary`
WHERE upper(country_region) = 'HUNGARY' AND date between '2021-10-01'
AND (select max(date) from `bigquery-public-data.covid19_jhu_csse.summary`)

AND FORMAT_DATE('%A', date) != 'Monday'
))

select *, FORMAT_DATE('%A', date_21) as nameofday from data_21
where new_cases_21 > 0
order by date_21 asc