Model updates for v4.0.0 and changes to population variants table and… #242
Conversation
|
The model updates text has now be okayed by James Sandy, with a few minor alterations that have been incorporated into this PR. |
|
|
||
| Sub-national population projections include four variant projections based on different assumptions about migration levels and trends. To provide a wider range of demographic scenarios for use in the model we create a custom set of sub-national variant projections using the [17 national variant projections produced by the ONS](https://www.ons.gov.uk/peoplepopulationandcommunity/populationandmigration/populationprojections/methodologies/variantnationalpopulationprojectionsfortheukandsubnationalpopulationprojectionsandhouseholdprojectionsforenglanduserguide). We do this by applying growth rates by age, sex, and year from the national variant projections to the base population for each LAD from the sub-national projections. | ||
|
|
||
| Where a national variant is based on the same set of assumptions as a sub-national variant, we prefer the sub-national version. The [complete set of projections](list_of_projections.qmd) available to the model is available in this documentation and includes 19 variants alongside the principal projection. | ||
| Where a national variant is based on the same set of assumptions as a sub-national variant, we prefer the sub-national version. The [complete set of projections](list_of_projections.qmd) available to the model is available in this documentation and includes 17 variants alongside the main ('migration category') projection. |
There was a problem hiding this comment.
as we are no longer using the national variants, is this redundant?
|
|
||
| ### Code restructure and new API | ||
|
|
||
| We have restructured our model code base significantly for this release and have also created a new Application Programme Interface (API) for the model. |
There was a problem hiding this comment.
the new-API is really just the old-API... there are some techincal reasons relating to switching internally (basically, newer features in Azure that required a complete re-deployment).
that, and the API is there purely from a data security perspective, it doesn't have any impact on the project otherwise.
So I probably wouldn't mention this.
There was a problem hiding this comment.
Agreed. If we're not mentioning new server then why mention new API
| These are 2022-based projections by single year of age and sex for Local Authorities. | ||
| The ONS have advised that, due to more recently updated migration data becoming available, their ‘migration variant’ should now be used as the new ‘principal projection’, as this is a better reflection of short-term population change. | ||
|
|
||
| The earliest baseline year that models can use based on this change is now 2023. |
There was a problem hiding this comment.
| The earliest baseline year that models can use based on this change is now 2023. | |
| The only baseline year that models can use based on this change is now 2023. |
|
If we are removing national projections then the info in this update is wrong and needs to be corrected. see https://github.com/The-Strategy-Unit/nhp_planning/issues/346#issuecomment-3043941297 for list of variants for v4.0 |
… text
Added model updates section for the population projections changes and API/code restructure. Also made relevant modifications to the text of the demographic modelling chapter to reflect the currently available subnational variants.
Draft until James S has had a chance to proof read the model updates section.
closes #239