Closed rsalzer closed 2 years ago
JU is informed. Thanks.
It is quite possible that JU's source on other days differs from the resource we scraped. We scrape what the cantons publish(ed). Establishing congruence retrospectively has its limits.
Well, congruence is quite easy, if the day before and the day after is the same ... it could just be corrected
Nah, the complete time series would have to be made congruent with the official source of the canton; we asked JU, what their source would be.
JU confirmed (via e-mail) an 'ncumul_deceased' of 151. Time series in https://github.com/openZH/covid_19/blob/master/fallzahlen_kanton_total_csv_v2/COVID19_Fallzahlen_Kanton_JU_total.csv is at 137. Where do we scrape the values of that variable from - via https://www.jura.ch/fr/Autorites/Coronavirus/Infos-Actualite/Statistiques-COVID/Evolution-des-cas-COVID-19-dans-le-Jura.html or ?
Ok, thats two different issues:
This here is probably just a typo: ncumul_deceased: 2021-02-07: 115 2021-02-08: 8 2021-02-09: 115
The number of deceased is summed up within the scraper from the table on the webpage. The dropping to 8 is a bug in the scraper, that I wasn't really able to spot/fix. I've created a bit a different approach to the current JU scraper in #1846 that isn't perfect too, but should cover things. I've just updated it and the test run produces 145 deceased (also with summing it up), so the data isn't matching the expected number, but is at least a bit closer...
This should finally be addressed with #1956
Only one day seems to be affected ... before and after the numbers are correct.