alan-turing-institute / daedalus

Daedalus is a novel dynamic spatial microsimulation pipeline that allows users to produce (custom) population projections for policy intervention analysis. Currently, it provides simulation utilities for the whole of the United Kingdom at the local authority (LA) level.
https://alan-turing-institute.github.io/daedalus/
MIT License
5 stars 1 forks source link

Make sure the LADs such as E09000001, E09000033 do not break on the pipeline, document this issue. #23

Closed crangelsmith closed 4 years ago

crangelsmith commented 4 years ago

This has been fixed on daedalous and the persistent data. Need to document the implication of this clearly on the README.

crangelsmith commented 4 years ago

Hi @kasra-hosseini, when you are working on the Daedalus README, could you add this text at the end of it?

===================

Warning

If you are planning to run the microsimulation pipeline on the LADs 'E09000001', 'E09000033', 'E06000052' and 'E06000053' beware that these are not treated independently on the rates as the rest of LADs. They are merged together in the following way:

You should be able to run on the single LAD files independently, but beware that the pipeline will be using the rates and total immigrated values for those LADs combined. The most appropriate way to deal with this is to run the microsimulation from a combined LAD starting file, instead of individually.

E.g for the LADs E09000001 and E09000033:

  1. Create a file named: ssm_E09000001+E09000033_MSOA11_ppp_2011.csv that contains the starting population from both E09000001 and E09000033.
  2. Run the pipeline in the following way: python scripts/run.py -c config/default_config.yaml --location E09000001+E09000033 --input_data_dir data --persistent_data_dir persistent_data --output_dir output

The equivalent should be done for 'E06000052' and 'E06000053' LADs.