Deces 2019 doesn't just contain 2019 deaths

[lfmcmillan]

Hi,
The deaths in Deces 2019.zip contain a range of death dates going back to 1935. This is visible in the deaths date format column and in the death dates txt column.

The Deces_2010_2018.zip file appears to contain a lot of deaths in 2009 and 2010 and a lot before than going back to 1800s, but none after 2010.

The Deces_2000_2009.zip file appears to contain a lot of deaths in 1999 and 2000, but none after mid-December 2000

Thanks
Louise

[scrouzet]

Hi Louise,

Some of these issues can be explained by the fact that the files follow the date of report by the INSEE, and not the date of death. For example, there could be a death that occurred in 1935 that actually get into INSEE data only in 2014. However this does not explain why the 2010:2018 would not contain any after 2010. This is odd.

In any case, this repo is not maintained anymore. If you are interested in these data, you can try directly to look on the INSEE website, they have improved a lot their sharing a few weeks after the COVID started (which is after I worked on this repo). I've seen that Baptiste Coulmont (for example here) is using these data to make analyses that are very close to what I aimed at here. You can also contact Jerome Hugueny, one of the contributor of this repo, who has also done very interesting analyses from these data.

Good luck,
Sébastien

[lfmcmillan]

Ah ok, thanks for the info Sébastien. I'm using the data for my data science teaching -- working with time series.