Merge pull request #48 from rcpch:lsoas

LSOA-LAD-addition

documentation/docs/developer/boundaries.md

@@ -0,0 +1,42 @@
+---
+title: Ordnance Survey and Boundaries
+author: Dr Simon Chapman
+---
+
+## Boundaries
+
+Healthcare often reports outcomes according to health geographies - for example at the level of Integrated Care Boards (ICBs) in England, but many like to have their data in administrative contexts also. Deprivation index, for example, which has a large impact on health, is not measured across health geographies, but across administrative ones.
+
+### Lower Layer Super Output Areas (LSOA)
+
+The index of multiple deprivation [for which RCPCH has another repository](https://github.com/rcpch/rcpch-census-platform) relies on census data to summarize different societal and population features to rank geographical units in order of most deprived to least deprived - the higher the score, the better off the region is. IMD uses Lower Layer Super Output Areas (LSOA) whose boundaries were last defined in 2011. These boundaries don't change much, but were updated last in 2021. They are not meant to change to allow longitudinal data to be gathered and be meaningful. LSOAs are chosen to be the way they are as they have population numbers as their common feature (1500 individuals or 650 households) - geographically distributed LSOAs cover a larger area, but in cities they are small as population numbers per unit area are larger.
+
+The last index of multiple of deprivation data for England was published in 2019 and used the 2011 LSOAs, since those were the latest boundaries. Since then, there has been the census of 2021, with the LSOA boundaries updated the same year. The decision on which boundaries will be included in the next iteration of IMD will be decided bythe Office for National Statistics (ONS) and the Ministry of Housing, Communities & Local Government (MHCLG) (now the Department for Levelling Up, Housing and Communities).
+
+### Local Authorities
+
+These are administrative geography boundaries. LSOAs conveniently fit neatly inside LAs but the LA boundaries change more frequently.
+
+**Local Authority Districts (LADs)** are a unit of local administration and statistical geography, and their structure varies across the country. Depending on the region, LADs can take different forms:
+
+#### Two-Tier Systems
+
+In some areas, LADs correspond to district councils, which operate alongside a county council that provides upper-tier services (e.g., Hertfordshire has multiple district councils, such as Watford Borough Council, each of which is an LAD).
+
+#### Unitary Authorities
+
+In other areas, LADs correspond to unitary authorities, which combine the responsibilities of both district and county councils into a single authority.
+
+#### Metropolitan Boroughs
+
+In metropolitan areas, LADs correspond to metropolitan boroughs (e.g., Manchester Metropolitan Borough), which provide local services. These boroughs are part of a wider metropolitan county (e.g., Greater Manchester) that handles strategic functions like transport and policing.
+
+##### London
+
+London is unique, with 32 London Boroughs serving as LADs, plus the City of London, which is its own LAD. The Greater London Authority (GLA) oversees strategic citywide functions but is not an LAD
+
+LAD boundaries change quite frequently. Each LAD has its own unique identifier and the boundaries and membership of LADs, which LSOAs they comprise, are updated by the ONS and published in the form of [look up tables](https://geoportal.statistics.gov.uk/documents/ons::local-authority-districts-counties-and-unitary-authorities-april-2023-map-in-the-uk/about?path=)
+
+## Implications for RCPCH NHS Organisations
+
+Although local authorities are not health geographies, local government has an interest in the health of the local population, and a responsibility for all functions of government that influence health (from housing, to green space, to transport etc), so RCPCH stores the relationship between the organisations that care for children and their local authority district.

documentation/docs/developer/shapes.md

@@ -30,4 +30,118 @@ RCPCH in the main uses the most generalised views - this is because more detail
 
 ### Boundary to IMD
 
-2011 LSOAs mapped to 2019 IMD data is a service fortunately already provided by [Consumer Data Research Centre](https://data.cdrc.ac.uk/dataset/index-multiple-deprivation-imd)
+2011 LSOAs mapped to 2019 IMD data is a service fortunately already provided by [Consumer Data Research Centre](https://data.cdrc.ac.uk/dataset/index-multiple-deprivation-imd)
+
+## Importing a .shp file
+
+1. Download the file from whichever source (using the LocalAuthorityDistrict model as an example)
+2. On the command line:
+    `python manage.py ogrinspect rcpch_nhs_organisations/hospitals/shape_files/Local_Authority_Districts_May_2024_Boundaries_UK_BUC/LAD_MAY_2024_UK_BUC.shp LocalAuthorityDistrict --srid 27700 --mapping --multi`
+    This will generate the code for the model (don't forget to import in `__init__.py` and add to admin):
+
+    ```python
+    class LocalAuthorityDistrict(models.Model):
+        lad24cd = models.CharField(max_length=9)
+        lad24nm = models.CharField(max_length=36)
+        lad24nmw = models.CharField(max_length=24)
+        bng_e = models.BigIntegerField()
+        bng_n = models.BigIntegerField()
+        long = models.FloatField()
+        lat = models.FloatField()
+        globalid = models.CharField(max_length=38)
+        geom = models.MultiPolygonField(srid=27700)
+
+
+    # Auto-generated `LayerMapping` dictionary for LocalAuthorityDistrict model
+    localauthoritydistrict_mapping = {
+        'lad24cd': 'LAD24CD',
+        'lad24nm': 'LAD24NM',
+        'lad24nmw': 'LAD24NMW',
+        'bng_e': 'BNG_E',
+        'bng_n': 'BNG_N',
+        'long': 'LONG',
+        'lat': 'LAT',
+        'globalid': 'GlobalID',
+        'geom': 'MULTIPOLYGON',
+    }
+    ```
+
+3. Create a migration for the new model (check previous examples, as RCPCH tend to add this as an abstract model, so this example diverges from actual practice for simplicity)
+    `python manage.py makemigrations`
+
+4. Create an empty migration
+    `python manage.py makemigrations hospitals --name seed_local_authority_districts_boundaries --empty`
+
+5. Create a custom function in the empty migration to import the .shp file geometry data using the layer map created earlier.
+
+    ```python
+    from django.db import migrations
+    from django.apps import apps as django_apps
+
+    import os
+    from django.contrib.gis.utils import LayerMapping
+
+    """
+    Local Authority Districts May 2024 Boundaries UK BUC
+    https://geoportal.statistics.gov.uk/search?q=BDY_LAD%202024&sort=Title%7Ctitle%7Casc
+    """
+
+    # Auto-generated `LayerMapping` dictionary for LocalAuthorityDistrict model
+    localauthoritydistrict_mapping = {
+        "lad24cd": "LAD24CD",
+        "lad24nm": "LAD24NM",
+        "lad24nmw": "LAD24NMW",
+        "bng_e": "BNG_E",
+        "bng_n": "BNG_N",
+        "long": "LONG",
+        "lat": "LAT",
+        "globalid": "GlobalID",
+        "geom": "MULTIPOLYGON",
+    }
+
+
+    # Boundary files
+
+    app_config = django_apps.get_app_config("hospitals")
+    app_path = app_config.path
+
+    Local_Authority_Districts_May_2024_Boundaries_UK_BUC = os.path.join(
+        app_path,
+        "shape_files",
+        "Local_Authority_Districts_May_2024_Boundaries_UK_BUC",
+        "LAD_MAY_2024_UK_BUC.shp",
+    )
+
+
+    def load(apps, schema_editor, verbose=True):
+        LocalAuthorityDistrict = apps.get_model("hospitals", "LocalAuthorityDistrict")
+        lm = LayerMapping(
+            LocalAuthorityDistrict,
+            Local_Authority_Districts_May_2024_Boundaries_UK_BUC,
+            localauthoritydistrict_mapping,
+            transform=False,
+            encoding="utf-8",
+        )
+        lm.save(strict=True, verbose=verbose)
+
+
+    class Migration(migrations.Migration):
+        dependencies = [
+            ("hospitals", "0012_localauthoritydistrict"),
+        ]
+
+        operations = [migrations.RunPython(load)]
+    ```
+
+6. Migrate the changes
+    `python manage.py migrate`
+7. If successful you should see:
+
+    ```console
+    ....
+    Saved: LocalAuthorityDistrict object (360)
+    Saved: LocalAuthorityDistrict object (361)
+    OK
+    ```
+
+8. This populates the model with the geometry files for mapping all the local authority boundaries, the names and codes. The next step is to hook this up to the other models.

documentation/mkdocs.yml

@@ -139,6 +139,8 @@ nav:
     - 'developer/models.md'
     - 'developer/seeding.md'
     - 'developer/database_maintenance.md'
+    - 'developer/shapes.md'
+    - 'developer/boundaries.md'
   - Contact:
     - 'contact/contact-rcpch.md'
   - Legal: