Add Scotland-specific calibration targets

changelog_entry.yaml

@@ -0,0 +1,4 @@
+- bump: minor
+  changes:
+    added:
+      - Scotland-specific calibration targets for children under 16 (NRS mid-year estimates) and households with 3+ children (Census 2022) to improve accuracy of Scottish policy analysis.

policyengine_uk_data/storage/demographics.csv

@@ -74,6 +74,8 @@ scotland_age_50_59,person-k,ons_age_sex_region,,,,,797,793,790,788,786,784,782,,
 scotland_age_60_69,person-k,ons_age_sex_region,,,,,691,702,714,727,741,755,767,,,
 scotland_age_70_79,person-k,ons_age_sex_region,,,,,506,520,534,546,557,565,575,,,
 scotland_age_80_89,person-k,ons_age_sex_region,,,,,274,282,290,297,302,307,312,,,
+scotland_children_under_16,person-k,nrs_mid_year_population_estimates,,,,,904,900,896,892,888,884,880,,,
+scotland_households_3plus_children,household-k,scotland_census_2022,,,,,82,82,82,82,82,82,82,,,
 south_east_age_0_9,person-k,ons_age_sex_region,,,,,1047,1044,1042,1039,1034,1028,1024,,,
 south_east_age_10_19,person-k,ons_age_sex_region,,,,,1122,1119,1117,1114,1108,1102,1097,,,
 south_east_age_20_29,person-k,ons_age_sex_region,,,,,1046,1062,1079,1093,1106,1114,1118,,,

policyengine_uk_data/utils/loss.py

@@ -263,6 +263,29 @@ def pe_count(*variables):
 
     df["ons/uk_population"] = household_from_person(age >= 0)
 
+    # Scotland-specific calibration targets
+    # Children under 16 in Scotland
+    # Source: NRS mid-year population estimates
+    # https://www.nrscotland.gov.uk/statistics-and-data/statistics/statistics-by-theme/population/population-estimates/mid-year-population-estimates
+    scotland_children_under_16 = (region == "SCOTLAND") & (age < 16)
+    df["ons/scotland_children_under_16"] = household_from_person(
+        scotland_children_under_16
+    )
+
+    # Households with 3+ children in Scotland
+    # Source: Scotland Census 2022 - Household composition
+    # https://www.scotlandscensus.gov.uk/census-results/at-a-glance/household-composition/
+    # Count children per household, filter to Scotland households with 3+
+    is_child = sim.calculate("is_child").values
+    children_per_household = household_from_person(is_child)
+    household_region = sim.calculate("region", map_to="household").values
+    scotland_3plus_children = (household_region == "SCOTLAND") & (
+        children_per_household >= 3
+    )
+    df["ons/scotland_households_3plus_children"] = (
+        scotland_3plus_children.astype(float)
+    )
+
     targets = (
         statistics[statistics.time_period == int(time_period)]
         .set_index("name")