1. Ethnicity:

[pbronka]

This sub-issue introduces a new ethnicity variable. See the parent issue for more details.

[pbronka]

To implement ethnicity:

• Get UKHLS and WAS data from the UKDS and generate input data for the model using these datasets and the do files from GitHub
• Load the "dot01" variable from the initial population csv files into the simulation:
• Define a 6-category ethnicity enum in the simulation, see e.g. https://github.com/centreformicrosimulation/SimPaths/blob/main/src/main/java/simpaths/model/enums/Gender.java or https://github.com/centreformicrosimulation/SimPaths/blob/main/src/main/java/simpaths/model/enums/Education.java
• Add "dot01" variable to the list of individual characteristics that are loaded into the database by modifying https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/data/Parameters.java#L99-L152
• Modify the SQL data parser class to load "dot01" variable similarly as gender is created from the dgn variable, see here https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/data/startingpop/DataParser.java#L143-L149. Make sure that the coding is consistent between the do file creating initial populations, the enum in the simulation, and the SQL parser.
• Define ethnicity variable in the Person class, see how gender is defined for an example https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/model/Person.java#L66. For ethnicity it will most likely be something like `@Enumerated(EnumType.STRING) private Ethnicity ethnicity; (Or call it "dot" to follow the EUROMOD naming convention)
• Add a getter and setter for the ethnicity variable
• Update the constructors in the Person class to include this variable, see how gender is handled by the copy constructor for example: https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/model/Person.java#L312
• Update the Person class constructor used to "create" children, see https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/model/Person.java#L240-L287 The way in which dehm_c3 for example is assigned based on mother's level of education can be followed to assign ethnicity from the mother too.
• Update the https://github.com/centreformicrosimulation/SimPaths/blob/e1e3c79b65bc704053898b3061a557f693e669c5/src/main/java/simpaths/model/Person.java#L1982 to add all the required cases corresponding to regressors in the Excel files in the input folder. Two options: aggregate from 6 categories up to 4 and 5 on the fly inside of these cases; or implement a conversion method inside of the ethnicity enum that will return a 4 or 5-category version of the variable (but note that the ethnicity entries are still required in the "DoublesVariables" (Person class) or "Regressors" (BenefitUnit class) just not implementing the logic of the conversion). (TODO: investigate the most appropriate way of implementing that conversion).
• Update the variable codebook to include the ethnicity variable

[dav-sonn]

This is done! As discussed, Ethnicity was recoded following the six categories specified in the parent issue and based on the original variable dot01. @igelstorm, we noticed that you had already inserted the variable in June (with a five-category specification) built upon the variable dot. Is it relevant to keep both categorisations, or can we give up the latter one? Were you using it somewhere specifically?
Thanks in advance

[andrewbaxter439]

I think financial distress is the only process so far using dot, so depends on all persons having a dot variable and the required Ethinicity{class} enums to call the five categories? An update to a six-category one will perhaps need re-estimation of financial distress regressions?

[andrewbaxter439]

I think #95 can be closed if combined into this one?