Old and new regions for cases presents in estimates - UK

[seabbs]

@kathsherratt could you flag the old depreciated UK regions that we should remove from the repository. At the moment they are shown on the website and in all downstream CSVs.

@joeHickson I think the best solution to this is to delete the folders for these regions in order to stop them appearing in the summaries.

[kathsherratt]

Thanks for noticing this Sam, sorry I missed it.

For cases, deaths, and admissions, we need to remove:

• Yorkshire and The Humber
• North East
• West Midlands
• East Midlands

For admissions only, we need to remove:

• United Kingdom

[seabbs]

@joeHickson could you clean these folders out from the production server? I think the Cornish internet is not ready for a 20Gb download and I don't have a VM with this loaded at the mo 😸

[seabbs]

This is getting fairly urgent. Can anyone look at this please?

I see multiple regions being updated for deaths in a way that I don't understand:

https://github.com/epiforecasts/covid-rt-estimates/tree/master/subnational/united-kingdom/deaths/national

@kathsherratt are there duplicate regions in the data source?

I have removed the UK for admissions.

[joeHickson]

Certainly the list at the point it's passed to epinow 2 is different:

2020-10-14 01:01:11 INFO Producing estimates for: East of England, England, London, Midlands, North East and Yorkshire, North West, Northern Ireland, Scotland, South East, South West, Wales, United Kingdom
2020-10-15 00:59:27 INFO Producing estimates for: East Midlands, East of England, England, London, North East, North West, Northern Ireland, Scotland, South East, South West, Wales, West Midlands, Yorkshire and The Humber, United Kingdom

We don't have logging of what locations were in the original data.

[joeHickson]

Is it possible the data source has changed?

[kathsherratt]

The data source seems fine, returning the correct geographies (4 nations + 7 NHS regions) with data present for the correct variables (cases_new, deaths_new, and hosp_new_blend).

data <- covidregionaldata::get_regional_data("uk", nhsregions = TRUE)
unique(data$region)

But it does look like the 2 "new" regions that are only brought in by using the nhsregions = TRUE argument (Midlands; North East & Yorkshire) haven't updated recently, at least for cases (last updated 2 days ago).

Has anything changed with the use of data_args in dataset-list.R? Could it be a change where the argument nhsregions = TRUE doesn't get passed through to covidregionaldata?

[joeHickson]

let see if that reoslves it

[joeHickson]

looks like some more bits might have gone astray with the problematic merge of #73