qingyuanzhao / bets.covid19 Goto Github PK

Apart from a few countries, the data we have are sparse. We always observe the confirmation date, but the other useful information (in particular, symptom onset) is usually not mentioned in news article (or withheld by the public agencies). This situation could become better if we can get volunteers to contribute and verify date, but eventually we need to think about how to deal with the missing dates.

Suggestions are welcomed.

What's the meanings of "Y" & "L" in column "Outside"?Thanks

Please register your interest by replying to this thread. Sharing your ideas would be most welcomed.

A perhaps unique feature of this outbreak is the extraordinary decision to lock down the city of Wuhan (and subsequently the Hubei Province) on 23rd of January. The decision to suspend all transportation into or out of Wuhan since 10am on January 23 was only announced at 2am. It is fair to say that the vast majority of people in Wuhan did not anticipate this decision when they fell asleep the previous day.

However, people were not totally unprepared. Speaking as someone who followed the news about this outbreak since the very first notice on December 31st, I can say that the overall tone changed quite dramatically after the confirmed cases sharply increase on January 19th and a celebrated Chinese epidemiologist confirmed human-to-human transmission on January 20th.

The challenge we are facing is how to accurately incorporate these into the analysis. Suggestions are welcomed.

We have several temporal features in the dataset: infection, international arrival, symptom onset, initial medical visit, hospitalization, case confimation. How can we model this dynamics? A problem specific to this data is every country has different levels of preparedness and disease control procedures.

Suggestions are welcomed.

Hi there. This is really interesting work.

I notice some aspects of this repo as an R package are non-standard.

Some (hopefully helpful) comments:

• The repo has a different name to the package. This will confuse inexperienced R users.

• Your package is called BETS. However there is already a package called BETS on CRAN. See here. Hence when users update thier package after yours is installed, using say update.packages(), the CRAN BETS package will replace your package because that has a higher version number than yours. Your best approach here is to rename your package to a name not used on CRAN. A convenient way to check available package names is to use

  available::available("your-proposed-pkg-name")

• Have you run R CMD check a.k.a. devtools::check() on this package? Currently running devtools::check() returns

  0 errors v | 2 warnings x | 3 notes x
  Error: R CMD check found WARNINGs
  Execution halted
  
  Exited with status 1.
  

  The rest of that output flags some of the non-standard issues:

  • Simple typo: in covid19_data there is a mismatch between Cases in the docs and the Case variable in the rda file.
  • The bayesian folder should not be at the top level. You need to exclude this from the package build in a .Rbuildignore file or move this folder to an appropriate subdirectory.

There are some other issues listed in the devtools::check() output. If you would like a PR fixing some of those, I am happy to try.

Best wishes
Tom

Suggestions by Cindy Chen in the Spreadsheet:

我发现北京卫健委每天都会以图标格式公布确诊患者的信息， 如果想做些关于效率的国别分析，这些数据都很clean的
http://wjw.beijing.gov.cn/xwzx_20031/wnxw/202002/t20200201_1622225.html
http://wjw.beijing.gov.cn/xwzx_20031/xwfb/202001/t20200131_1622025.html

上海卫建委也每天提供图表，只是关注的点不太一样
http://wsjkw.sh.gov.cn/xwfb/20200202/f5b6ce31678e43dca76c09d93ec4142b.html

合肥的报道也做的非常好，和你用的列都是一样的，我想这组数据可以用于做对照表http://www.bjnews.com.cn/news/2020/01/27/680324.html

If someone volunteers to add the data from Hefei, I can give you permission to edit the Spreadsheet.