Move transform logic around

This commit is contained in:
Andras Schmelczer 2026-01-31 12:48:29 +00:00
parent e1b38a1b95
commit 38b0cf1ea1
14 changed files with 1073 additions and 336 deletions

View file

@ -1,38 +1,136 @@
version: '3'
vars:
DATA_DIR: data
ARCGIS_OUTPUT: "{{.DATA_DIR}}/arcgis_data.parquet"
PRICE_PAID_OUTPUT: "{{.DATA_DIR}}/price-paid-complete.parquet"
IOD_OUTPUT: "{{.DATA_DIR}}/IoD2025_Scores.parquet"
POIS_RAW_OUTPUT: "{{.DATA_DIR}}/uk_pois.parquet"
POIS_FILTERED_OUTPUT: "{{.DATA_DIR}}/filtered_uk_pois.parquet"
POI_PROXIMITY_OUTPUT: "{{.DATA_DIR}}/poi_proximity.parquet"
EPC_PP_OUTPUT: "{{.DATA_DIR}}/epc_pp.parquet"
WIDE_OUTPUT: "{{.DATA_DIR}}/wide.parquet"
EPC_CSV: "{{.DATA_DIR}}/epc/certificates.csv"
JOURNEY_TIMES: "{{.DATA_DIR}}/journey_times_bank_checkpoint.parquet"
tasks:
install:
desc: Install dependencies, generate client, and download data
desc: Install dependencies
cmds:
- uv sync
- cd frontend && npm install
download:
desc: Download data
deps:
- install
download:arcgis:
desc: Download and convert ArcGIS postcode data
sources:
- pipeline/download/arcgis.py
generates:
- "{{.ARCGIS_OUTPUT}}"
cmds:
- uv run -m pipeline.download.arcgis
- uv run -m pipeline.download.pois
- uv run -m pipeline.download.deprivation_data
- uv run -m pipeline.download.price_paid
- uv run python -m pipeline.download.arcgis --output {{.ARCGIS_OUTPUT}}
pipeline:
desc: Run data processing pipeline
deps:
- download
download:price-paid:
desc: Download and convert Land Registry price-paid data
sources:
- pipeline/download/price_paid.py
generates:
- "{{.PRICE_PAID_OUTPUT}}"
cmds:
- uv run python -m pipeline.run
- uv run python -m pipeline.download.price_paid --output {{.PRICE_PAID_OUTPUT}}
download:deprivation:
desc: Download and convert Index of Deprivation data
sources:
- pipeline/download/deprivation_data.py
generates:
- "{{.IOD_OUTPUT}}"
cmds:
- uv run python -m pipeline.download.deprivation_data --output {{.IOD_OUTPUT}}
download:pois:
desc: Download and extract POIs from OpenStreetMap
sources:
- pipeline/download/pois.py
generates:
- "{{.POIS_RAW_OUTPUT}}"
cmds:
- uv run python -m pipeline.download.pois --output {{.POIS_RAW_OUTPUT}}
transform:pois:
desc: Transform raw POIs to filtered version with friendly names
deps:
- download:pois
sources:
- pipeline/transform/transform_poi.py
- "{{.POIS_RAW_OUTPUT}}"
generates:
- "{{.POIS_FILTERED_OUTPUT}}"
cmds:
- uv run python -m pipeline.transform.transform_poi --input {{.POIS_RAW_OUTPUT}} --output {{.POIS_FILTERED_OUTPUT}}
transform:epc-pp:
desc: Fuzzy join EPC and Price Paid data
deps:
- download:price-paid
sources:
- pipeline/transform/join_epc_pp.py
- pipeline/utils/fuzzy_join.py
- "{{.PRICE_PAID_OUTPUT}}"
- "{{.EPC_CSV}}"
generates:
- "{{.EPC_PP_OUTPUT}}"
cmds:
- uv run python -m pipeline.transform.join_epc_pp --epc {{.EPC_CSV}} --price-paid {{.PRICE_PAID_OUTPUT}} --output {{.EPC_PP_OUTPUT}}
transform:poi-proximity:
desc: Compute POI proximity counts per postcode
deps:
- download:arcgis
- transform:pois
sources:
- pipeline/transform/poi_proximity.py
- pipeline/utils/poi_counts.py
- "{{.ARCGIS_OUTPUT}}"
- "{{.POIS_FILTERED_OUTPUT}}"
generates:
- "{{.POI_PROXIMITY_OUTPUT}}"
cmds:
- uv run python -m pipeline.transform.poi_proximity --arcgis {{.ARCGIS_OUTPUT}} --pois {{.POIS_FILTERED_OUTPUT}} --output {{.POI_PROXIMITY_OUTPUT}}
transform:wide:
desc: Build wide property dataframe with all joins
deps:
- join:epc-pp
- download:arcgis
- download:deprivation
- transform:poi-proximity
sources:
- pipeline/transform/merge.py
- "{{.EPC_PP_OUTPUT}}"
- "{{.ARCGIS_OUTPUT}}"
- "{{.IOD_OUTPUT}}"
- "{{.POI_PROXIMITY_OUTPUT}}"
generates:
- "{{.WIDE_OUTPUT}}"
cmds:
- uv run python -m pipeline.transform.merge --epc-pp {{.EPC_PP_OUTPUT}} --arcgis {{.ARCGIS_OUTPUT}} --iod {{.IOD_OUTPUT}} --poi-proximity {{.POI_PROXIMITY_OUTPUT}} --journey-times {{.JOURNEY_TIMES}} --output {{.WIDE_OUTPUT}}
prepare:
desc: Prepare the application (install, download data, run pipeline)
deps:
- pipeline
- transform:wide
test:
cmds:
- uv run -m pipeline.utils.test_fuzzy_join
- uv run pytest pipeline/utils/test_haversine.py
- uv run pytest pipeline/utils/test_poi_counts.py
server:
desc: Run FastAPI backend on port 8001
desc: Run Rust backend on port 8001
dir: server-rs
cmds:
- uv run fastapi dev server/main.py --port 8001
- cargo run --release -- {{.WIDE_OUTPUT}}
frontend:
desc: Run frontend dev server on port 3030 (proxies /api to :8001)
@ -46,16 +144,13 @@ tasks:
cmds:
- npm run build
prod:
desc: Run production server (serves built frontend)
cmds:
- uv run fastapi run server/main.py --port 8001
lint:
desc: Lint all code (Python and TypeScript)
desc: Lint all code (Python, TypeScript, and Rust)
cmds:
- task: lint:python
- task: lint:frontend
- task: lint:rust
lint:python:
desc: Lint Python code with ruff
@ -69,11 +164,19 @@ tasks:
- npm run lint
- npm run format:check
lint:rust:
desc: Lint Rust code with clippy and check formatting
dir: server-rs
cmds:
- cargo clippy -- -D warnings
- cargo fmt --check
format:
desc: Format all code (Python and TypeScript)
desc: Format all code (Python, TypeScript, and Rust)
cmds:
- task: format:python
- task: format:frontend
- task: format:rust
format:python:
desc: Format Python code with ruff
@ -88,6 +191,12 @@ tasks:
- npm run lint:fix
- npm run format
format:rust:
desc: Format Rust code with cargo fmt
dir: server-rs
cmds:
- cargo fmt
check:
desc: Run all checks (lint, typecheck, build)
cmds:

View file

@ -1,22 +0,0 @@
from abc import ABC, abstractmethod
import polars as pl
class DataSource(ABC):
"""Base class for all data sources."""
@property
@abstractmethod
def name(self) -> str:
"""Unique identifier for this data source."""
pass
@abstractmethod
def load(self) -> pl.LazyFrame:
"""Load raw data as LazyFrame."""
pass
@abstractmethod
def process(self, postcodes: pl.LazyFrame) -> pl.LazyFrame:
"""Process and join with postcode coordinates."""
pass

View file

@ -1,12 +0,0 @@
"""Shared configuration for the pipeline and server."""
from pathlib import Path
DATA_DIR = Path(__file__).parent.parent / "data_sources"
PROCESSED_DIR = DATA_DIR / "processed"
AGGREGATES_DIR = PROCESSED_DIR / "aggregates"
# H3 resolutions to generate and serve
# https://h3geo.org/docs/core-library/restable/#average-area-in-m2
H3_RESOLUTIONS = [7, 8, 9, 10, 11]
DEFAULT_H3_RESOLUTION = 8

View file

@ -8,17 +8,35 @@ import osmium
import polars as pl
from tqdm import tqdm
from .config import (
BATCH_SIZE,
GEOFABRIK_GB_URL,
MIN_OCCURENCE_COUNT,
POI_TAG_KEYS,
UK_BBOX_EAST,
UK_BBOX_NORTH,
UK_BBOX_SOUTH,
UK_BBOX_WEST,
from pathlib import Path
BATCH_SIZE = 50_000
MIN_OCCURENCE_COUNT = 20
GEOFABRIK_GB_URL = (
"https://download.geofabrik.de/europe/great-britain-latest.osm.pbf"
)
UK_BBOX_WEST = -7.57
UK_BBOX_SOUTH = 49.96
UK_BBOX_EAST = 1.68
UK_BBOX_NORTH = 58.64
POI_TAG_KEYS: list[str] = [
"amenity",
"building",
"craft",
"emergency",
"healthcare",
"leisure",
"office",
"shop",
"tourism",
"public_transport",
]
def download_pbf(pbf_file: Path) -> None:
@ -144,10 +162,9 @@ def main() -> None:
)
df = df.join(valid_categories.select("category"), on="category", how="semi")
args.output.parent.mkdir(parents=True, exist_ok=True)
print(f"Total POIs: {handler.poi_count:,}")
df.sink_parquet(args.output)
print(f"Saved to {args.output}")
print(f"Total POIs: {handler.poi_count:,}")
if __name__ == "__main__":

View file

@ -1,32 +0,0 @@
from pathlib import Path
DATA_DIR = Path("./data_sources")
GB_PBF_FILE = DATA_DIR / "great-britain-latest.osm.pbf"
OUTPUT_FILE = DATA_DIR / "uk_pois.parquet"
BATCH_SIZE = 50_000
MIN_OCCURENCE_COUNT = 20
GEOFABRIK_GB_URL = (
"https://download.geofabrik.de/europe/great-britain-latest.osm.pbf"
)
UK_BBOX_WEST = -7.57
UK_BBOX_SOUTH = 49.96
UK_BBOX_EAST = 1.68
UK_BBOX_NORTH = 58.64
POI_TAG_KEYS: list[str] = [
"amenity",
"building",
"craft",
"emergency",
"healthcare",
"leisure",
"office",
"shop",
"tourism",
"public_transport",
]

View file

@ -1,85 +0,0 @@
import polars as pl
from .fuzzy_join import fuzzy_join_on_postcode
pl.Config.set_tbl_cols(-1)
epc = pl.scan_csv('data_sources/epc/certificates.csv').select(
pl.col('ADDRESS').alias('epc_address'),
'POSTCODE',
'CURRENT_ENERGY_RATING',
'POTENTIAL_ENERGY_RATING',
pl.col('PROPERTY_TYPE').alias('epc_property_type'),
'BUILT_FORM',
'INSPECTION_DATE',
'TOTAL_FLOOR_AREA',
'NUMBER_HABITABLE_ROOMS',
'FLOOR_HEIGHT',
'CONSTRUCTION_AGE_BAND'
).filter(pl.col('epc_address').is_not_null()).sort('INSPECTION_DATE', descending=True).group_by('epc_address', 'POSTCODE').first()
print("EPC dataset")
print(epc.head().collect())
# https://www.gov.uk/guidance/about-the-price-paid-data
property_type_map = {"D": "Detached", "S": "Semi-Detached", "T": "Terraced", "F": "Flats/Maisonettes", "O": "Other"}
duration_map = {"F": "Freehold", "L": "Leasehold"}
price_paid = (pl.scan_parquet('data_sources/pp-complete.parquet').select(
"price",
"date_of_transfer",
pl.col('property_type').alias("pp_property_type").replace(property_type_map),
"postcode",
'paon',
'saon',
'street',
'locality',
'town_city',
pl.col('duration').replace(duration_map)
)
.filter(pl.col('pp_property_type') != 'Other').with_columns(
pl.concat_str(
[pl.col('saon'), pl.col('paon'), pl.col('street')],
separator=' ',
ignore_nulls=True,
).alias('pp_address'),
)
.sort('date_of_transfer')
.group_by('pp_address', 'postcode', maintain_order=True)
.agg(
pl.struct(
pl.col('date_of_transfer').dt.year().alias('year'),
'price',
).alias('historical_prices'),
pl.col('pp_property_type').last(),
pl.col('duration').last(),
pl.col('price').last().alias('latest_price'),
pl.col('date_of_transfer').last(),
)
).filter(pl.col('pp_address').is_not_null())
print("Price paid dataset")
print(price_paid.head().collect())
joined = fuzzy_join_on_postcode(
left=price_paid,
right=epc,
left_address_col='pp_address',
right_address_col='epc_address',
left_postcode_col='postcode',
right_postcode_col='POSTCODE',
).drop('POSTCODE').collect()
matched = joined.filter(pl.col('epc_address').is_not_null() & pl.col('pp_address').is_not_null())
total = joined.height
print(f"Unique properties: {total}")
print(f"Matched: {matched.height} ({100 * matched.height / total:.1f}%)")
print(f"Unmatched: {total - matched.height}")
matched = matched.rename({col: col.lower() for col in joined.columns})
print(matched.head())
matched.write_parquet('data_sources/processed/epc_pp.parquet')

View file

@ -1,6 +0,0 @@
"""Pipeline CLI to process property data with H3 spatial indexing."""
from pipeline.wide import run
if __name__ == "__main__":
run()

View file

@ -0,0 +1,98 @@
import argparse
import polars as pl
from pathlib import Path
from ..utils import fuzzy_join_on_postcode
pl.Config.set_tbl_cols(-1)
def main():
parser = argparse.ArgumentParser(description="Fuzzy join EPC and Price Paid data")
parser.add_argument("--epc", type=Path, required=True, help="EPC certificates CSV file")
parser.add_argument("--price-paid", type=Path, required=True, help="Price paid parquet file")
parser.add_argument("--output", type=Path, required=True, help="Output parquet file path")
args = parser.parse_args()
epc = pl.scan_csv(args.epc).select(
pl.col('ADDRESS').alias('epc_address'),
'POSTCODE',
'CURRENT_ENERGY_RATING',
'POTENTIAL_ENERGY_RATING',
pl.col('PROPERTY_TYPE').alias('epc_property_type'),
'BUILT_FORM',
'INSPECTION_DATE',
'TOTAL_FLOOR_AREA',
'NUMBER_HABITABLE_ROOMS',
'FLOOR_HEIGHT',
'CONSTRUCTION_AGE_BAND'
).filter(pl.col('epc_address').is_not_null()).sort('INSPECTION_DATE', descending=True).group_by('epc_address', 'POSTCODE').first()
print("EPC dataset")
print(epc.head().collect())
# https://www.gov.uk/guidance/about-the-price-paid-data
property_type_map = {"D": "Detached", "S": "Semi-Detached", "T": "Terraced", "F": "Flats/Maisonettes", "O": "Other"}
duration_map = {"F": "Freehold", "L": "Leasehold"}
price_paid = (pl.scan_parquet(args.price_paid).select(
"price",
"date_of_transfer",
pl.col('property_type').alias("pp_property_type").replace(property_type_map),
"postcode",
'paon',
'saon',
'street',
'locality',
'town_city',
pl.col('duration').replace(duration_map)
)
.filter(pl.col('pp_property_type') != 'Other').with_columns(
pl.concat_str(
[pl.col('saon'), pl.col('paon'), pl.col('street')],
separator=' ',
ignore_nulls=True,
).alias('pp_address'),
)
.sort('date_of_transfer')
.group_by('pp_address', 'postcode', maintain_order=True)
.agg(
pl.struct(
pl.col('date_of_transfer').dt.year().alias('year'),
'price',
).alias('historical_prices'),
pl.col('pp_property_type').last(),
pl.col('duration').last(),
pl.col('price').last().alias('latest_price'),
pl.col('date_of_transfer').last(),
)
).filter(pl.col('pp_address').is_not_null())
print("Price paid dataset")
print(price_paid.head().collect())
joined = fuzzy_join_on_postcode(
left=price_paid,
right=epc,
left_address_col='pp_address',
right_address_col='epc_address',
left_postcode_col='postcode',
right_postcode_col='POSTCODE',
).drop('POSTCODE').collect()
matched = joined.filter(pl.col('epc_address').is_not_null() & pl.col('pp_address').is_not_null())
total = joined.height
print(f"Unique properties: {total}")
print(f"Matched: {matched.height} ({100 * matched.height / total:.1f}%)")
print(f"Unmatched: {total - matched.height}")
matched = matched.rename({col: col.lower() for col in joined.columns})
print(matched.head())
matched.write_parquet(args.output)
print(f"Wrote {args.output}")
if __name__ == "__main__":
main()

127
pipeline/transform/merge.py Normal file
View file

@ -0,0 +1,127 @@
import argparse
import polars as pl
from pathlib import Path
def _build_wide(
epc_pp_path: Path,
arcgis_path: Path,
iod_path: Path | None,
poi_proximity_path: Path | None,
journey_times_path: Path | None,
) -> pl.DataFrame:
"""Build the wide dataframe by joining epc_pp with all auxiliary data."""
print("Loading epc_pp...")
wide = pl.read_parquet(epc_pp_path)
print(f" {wide.shape[0]:,} rows, {wide.estimated_size('mb'):.1f} MB")
# GPS coordinates + LSOA from ArcGIS
print("Joining GPS coordinates...")
arcgis = pl.read_parquet(arcgis_path).select(
pl.col("pcds").alias("postcode"),
"lat",
pl.col("long").alias("lon"),
"lsoa21",
)
wide = wide.join(arcgis, on="postcode", how="inner")
print(f" {wide.shape[0]:,} rows after GPS join, {wide.estimated_size('mb'):.1f} MB")
# Journey times (optional)
if journey_times_path and journey_times_path.exists():
print("Joining journey times...")
journey_times = pl.read_parquet(journey_times_path).select(
"postcode",
"public_transport_easy_minutes",
"public_transport_quick_minutes",
"cycling_minutes",
)
wide = wide.join(journey_times, on="postcode", how="left")
print(f" {wide.estimated_size('mb'):.1f} MB after journey times")
# Index of Deprivation
if iod_path and iod_path.exists():
print("Joining IoD scores...")
iod = pl.read_parquet(iod_path)
wide = wide.join(
iod, left_on="lsoa21", right_on="LSOA code (2021)", how="left"
)
print(f" {wide.estimated_size('mb'):.1f} MB after IoD")
# POI proximity counts (pre-computed per postcode)
if poi_proximity_path and poi_proximity_path.exists():
print("Joining POI proximity counts...")
poi_counts = pl.read_parquet(poi_proximity_path)
wide = wide.join(poi_counts, on="postcode", how="left")
print(f" {wide.estimated_size('mb'):.1f} MB after POI counts")
# Convert construction_age_band to numeric year
if "construction_age_band" in wide.columns:
wide = wide.with_columns(
pl.col("construction_age_band")
.str.replace("England and Wales: ", "")
.str.replace(" onwards", "")
.str.extract(r"(\d{4})", 1)
.cast(pl.UInt16, strict=False)
.alias("construction_age_band"),
)
# Derived columns
wide = wide.with_columns(
(pl.col("latest_price") / pl.col("total_floor_area")).alias("Price per sqm"),
).drop(
'date_of_transfer',
'inspection_date',
'floor_height',
'lsoa21',
'LSOA code (2021)',
'Local Authority District code (2024)',
'Local Authority District name (2024)',
'imd_score',
'housing_barriers_score',
'idaci_score',
'idaopi_score',
'children_young_people_score',
'adult_skills_score',
'geographical_barriers_score',
'wider_barriers_score',
).rename({
'construction_age_band': "Approximate construction age",
"income_score": "Income Score (rate)",
"employment_score": "Employment Score (rate)",
"education_score": "Education, Skills and Training Score",
"health_score": "Health Deprivation and Disability Score",
"crime_score": "Crime Score",
})
return wide
def main():
parser = argparse.ArgumentParser(description="Build wide property dataframe with all joins")
parser.add_argument("--epc-pp", type=Path, required=True, help="EPC-Price Paid joined parquet file")
parser.add_argument("--arcgis", type=Path, required=True, help="ArcGIS postcode data parquet file")
parser.add_argument("--iod", type=Path, help="Index of Deprivation parquet file (optional)")
parser.add_argument("--poi-proximity", type=Path, help="POI proximity counts parquet file (optional)")
parser.add_argument("--journey-times", type=Path, help="Journey times parquet file (optional)")
parser.add_argument("--output", type=Path, required=True, help="Output parquet file path")
args = parser.parse_args()
wide = _build_wide(
epc_pp_path=args.epc_pp,
arcgis_path=args.arcgis,
iod_path=args.iod,
poi_proximity_path=args.poi_proximity,
journey_times_path=args.journey_times,
)
print(f"Columns: {wide.columns}")
print(f"Rows: {wide.height}")
wide.write_parquet(args.output)
size_mb = args.output.stat().st_size / (1024 * 1024)
print(f"Wrote {args.output} ({size_mb:.1f} MB)")
if __name__ == "__main__":
main()

View file

@ -0,0 +1,42 @@
"""Compute POI proximity counts per postcode from ArcGIS + filtered POIs."""
import argparse
from pathlib import Path
import polars as pl
from pipeline.utils.poi_counts import _count_pois_per_postcode
def main():
parser = argparse.ArgumentParser(
description="Count POIs within radius per postcode"
)
parser.add_argument(
"--arcgis", type=Path, required=True, help="ArcGIS postcode parquet"
)
parser.add_argument(
"--pois", type=Path, required=True, help="Filtered POIs parquet"
)
parser.add_argument(
"--output", type=Path, required=True, help="Output parquet path"
)
args = parser.parse_args()
postcodes = pl.read_parquet(args.arcgis).select(
pl.col("pcds").alias("postcode"),
"lat",
pl.col("long").alias("lon"),
)
pois = pl.read_parquet(args.pois)
result = _count_pois_per_postcode(postcodes, pois, radius_km=2)
result.write_parquet(args.output)
size_mb = args.output.stat().st_size / (1024 * 1024)
print(f"Wrote {args.output} ({size_mb:.1f} MB)")
if __name__ == "__main__":
main()

View file

@ -0,0 +1,644 @@
import argparse
from pathlib import Path
import polars as pl
DROP_CATEGORIES = {
"amenity/advice",
"amenity/atm",
"amenity/bbq",
"amenity/bench",
"amenity/bicycle_parking",
"amenity/clock",
"amenity/fixme",
"amenity/grit_bin",
"amenity/hunting_stand",
"amenity/motorcycle_parking",
"amenity/notice_board",
"amenity/parking",
"amenity/parking_entrance",
"amenity/parking_space",
"amenity/post_box",
"amenity/telephone",
"amenity/toilets",
"amenity/vacuum_cleaner",
"amenity/waste_basket",
"building/air_shaft",
"building/apartments",
"building/detached",
"building/entrance",
"building/entry",
"building/garage",
"building/garages",
"building/house",
"building/hut",
"building/no",
"building/office",
"building/public",
"building/residential",
"building/roof",
"building/shed",
"building/terrace",
"building/yes",
"emergency/access_point",
"emergency/ambulance_station",
"emergency/assembly_point",
"emergency/bleed_control_kit",
"emergency/defibrillator",
"emergency/designated",
"emergency/dry_riser_inlet",
"emergency/emergency_ward_entrance",
"emergency/fire_alarm_box",
"emergency/fire_extinguisher",
"emergency/fire_hydrant",
"emergency/fire_service_inlet",
"emergency/first_aid_kit",
"emergency/life_ring",
"emergency/lifeguard",
"emergency/no",
"emergency/phone",
"emergency/rescue_equipment",
"emergency/siren",
"emergency/throw_bag",
"emergency/water_rescue",
"emergency/yes",
"leisure/firepit",
"leisure/fishing",
"leisure/picnic_table",
"office/company",
"office/yes",
"tourism/apartment",
"tourism/apartments",
"tourism/camp_pitch",
"tourism/information",
"tourism/village_sign",
"tourism/yes",
}
# (friendly_name, emoji) for every category we keep
CATEGORY_MAP: dict[str, tuple[str, str]] = {
# amenity
"amenity/animal_boarding": ("Animal Boarding", "🐾"),
"amenity/animal_breeding": ("Animal Breeding", "🐣"),
"amenity/animal_shelter": ("Animal Shelter", "🏠"),
"amenity/arts_centre": ("Arts Centre", "🎨"),
"amenity/bank": ("Bank", "🏦"),
"amenity/bar": ("Bar", "🍸"),
"amenity/bicycle_rental": ("Bike Rental", "🚲"),
"amenity/bicycle_repair_station": ("Bike Repair", "🔧"),
"amenity/binoculars": ("Public Binoculars", "🔭"),
"amenity/boat_rental": ("Boat Rental", ""),
"amenity/boat_storage": ("Boat Storage", "🚢"),
"amenity/boot_scraper": ("Boot Scraper", "🥾"),
"amenity/bureau_de_change": ("Currency Exchange", "💱"),
"amenity/bus_station": ("Bus Station", "🚌"),
"amenity/cafe": ("Café", ""),
"amenity/car_rental": ("Car Rental", "🚗"),
"amenity/car_sharing": ("Car Sharing", "🚙"),
"amenity/car_wash": ("Car Wash", "🧽"),
"amenity/care_home": ("Care Home", "🏥"),
"amenity/casino": ("Casino", "🎰"),
"amenity/charging_station": ("EV Charging", "🔌"),
"amenity/check_in": ("Check-In Point", ""),
"amenity/childcare": ("Childcare", "👶"),
"amenity/cinema": ("Cinema", "🎬"),
"amenity/clinic": ("Clinic", "🩺"),
"amenity/club": ("Club", "🏛️"),
"amenity/college": ("College", "🎓"),
"amenity/community_centre": ("Community Centre", "🤝"),
"amenity/compressed_air": ("Compressed Air", "💨"),
"amenity/conference_centre": ("Conference Centre", "📋"),
"amenity/courthouse": ("Courthouse", "⚖️"),
"amenity/coworking_space": ("Co-working Space", "💻"),
"amenity/crematorium": ("Crematorium", "🕯️"),
"amenity/dancing_school": ("Dance School", "💃"),
"amenity/dentist": ("Dentist", "🦷"),
"amenity/doctors": ("Doctor", "👨‍⚕️"),
"amenity/dojo": ("Dojo", "🥋"),
"amenity/donation_box": ("Donation Box", "📦"),
"amenity/dressing_room": ("Dressing Room", "👗"),
"amenity/drinking_water": ("Drinking Water", "🚰"),
"amenity/driving_school": ("Driving School", "🚦"),
"amenity/escooter_rental": ("E-Scooter Rental", "🛴"),
"amenity/events_venue": ("Events Venue", "🎪"),
"amenity/fast_food": ("Fast Food", "🍔"),
"amenity/feeding_place": ("Feeding Place", "🍽️"),
"amenity/ferry_terminal": ("Ferry Terminal", "⛴️"),
"amenity/fire_station": ("Fire Station", "🚒"),
"amenity/food_court": ("Food Court", "🍴"),
"amenity/fountain": ("Fountain", ""),
"amenity/fuel": ("Fuel Station", ""),
"amenity/gambling": ("Gambling", "🎲"),
"amenity/grave_yard": ("Graveyard", "🪦"),
"amenity/hall": ("Hall", "🏛️"),
"amenity/hookah_lounge": ("Hookah Lounge", "💨"),
"amenity/hospital": ("Hospital", "🏥"),
"amenity/ice_cream": ("Ice Cream", "🍦"),
"amenity/internet_cafe": ("Internet Café", "🌐"),
"amenity/kick-scooter_rental": ("Kick Scooter Rental", "🛴"),
"amenity/kindergarten": ("Kindergarten", "💒"),
"amenity/language_school": ("Language School", "🗣️"),
"amenity/letter_box": ("Letter Box", "📮"),
"amenity/library": ("Library", "📚"),
"amenity/loading_dock": ("Loading Dock", "📥"),
"amenity/lounge": ("Lounge", "🛋️"),
"amenity/lounger": ("Public Lounger", "🪑"),
"amenity/marketplace": ("Market", "🛒"),
"amenity/money_transfer": ("Money Transfer", "💸"),
"amenity/mounting_block": ("Mounting Block", "🐴"),
"amenity/music_school": ("Music School", "🎵"),
"amenity/music_venue": ("Music Venue", "🎶"),
"amenity/nightclub": ("Nightclub", "🪩"),
"amenity/nursing_home": ("Nursing Home", "🏠"),
"amenity/parcel_locker": ("Parcel Locker", "📦"),
"amenity/payment_terminal": ("Payment Terminal", "💳"),
"amenity/pharmacy": ("Pharmacy", "💊"),
"amenity/photo_booth": ("Photo Booth", "📸"),
"amenity/piano": ("Public Piano", "🎹"),
"amenity/place_of_worship": ("Place of Worship", ""),
"amenity/police": ("Police Station", "🚔"),
"amenity/post_depot": ("Post Depot", "📬"),
"amenity/post_office": ("Post Office", "🏤"),
"amenity/prep_school": ("Prep School", "📖"),
"amenity/pub": ("Pub", "🍺"),
"amenity/public_bookcase": ("Public Bookcase", "📕"),
"amenity/public_building": ("Public Building", "🏢"),
"amenity/reception_desk": ("Reception Desk", "🛎️"),
"amenity/recycling": ("Recycling", "♻️"),
"amenity/restaurant": ("Restaurant", "🍽️"),
"amenity/sanitary_dump_station": ("Sanitary Dump Station", "🚿"),
"amenity/school": ("School", "🏫"),
"amenity/scout_hut": ("Scout Hut", "⚜️"),
"amenity/shelter": ("Shelter", "🛖"),
"amenity/shower": ("Public Shower", "🚿"),
"amenity/smoking_area": ("Smoking Area", "🚬"),
"amenity/social_centre": ("Social Centre", "🏘️"),
"amenity/social_club": ("Social Club", "🤝"),
"amenity/social_facility": ("Social Facility", "🫂"),
"amenity/stripclub": ("Strip Club", "🔞"),
"amenity/studio": ("Studio", "🎙️"),
"amenity/table": ("Public Table", "🪑"),
"amenity/taxi": ("Taxi Stand", "🚕"),
"amenity/telescope": ("Public Telescope", "🔭"),
"amenity/theatre": ("Theatre", "🎭"),
"amenity/ticket_validator": ("Ticket Validator", "🎫"),
"amenity/townhall": ("Town Hall", "🏛️"),
"amenity/training": ("Training Centre", "📝"),
"amenity/trolley_bay": ("Trolley Bay", "🛒"),
"amenity/university": ("University", "🏫"),
"amenity/vehicle_inspection": ("Vehicle Inspection", "🔍"),
"amenity/vending_machine": ("Vending Machine", "🏧"),
"amenity/veterinary": ("Vet", "🐕"),
"amenity/washing_machine": ("Washing Machine", "🧺"),
"amenity/washingline": ("Washing Line", "👕"),
"amenity/waste_disposal": ("Waste Disposal", "🗑️"),
"amenity/waste_transfer_station": ("Waste Transfer Station", "🚛"),
"amenity/water_point": ("Water Point", "💧"),
"amenity/watering_place": ("Watering Place", "🚰"),
"amenity/weighbridge": ("Weighbridge", "⚖️"),
# building
"building/barn": ("Barn", "🏚️"),
"building/bunker": ("Bunker", "🏗️"),
"building/chapel": ("Chapel", ""),
"building/church": ("Church", ""),
"building/commercial": ("Commercial Building", "🏬"),
"building/construction": ("Construction Site", "🚧"),
"building/farm": ("Farmhouse", "🌾"),
"building/greenhouse": ("Greenhouse", "🌿"),
"building/industrial": ("Industrial Building", "🏭"),
"building/kiosk": ("Kiosk", "🏪"),
"building/retail": ("Retail Building", "🏬"),
"building/ruins": ("Ruins", "🏚️"),
"building/school": ("School Building", "🏫"),
"building/semidetached_house": ("Semi-Detached House", "🏠"),
"building/service": ("Service Building", "🔧"),
"building/university": ("University Building", "🎓"),
"building/warehouse": ("Warehouse", "🏭"),
# craft
"craft/agricultural_engines": ("Agricultural Engines", "🚜"),
"craft/atelier": ("Atelier", "🎨"),
"craft/blacksmith": ("Blacksmith", "🔨"),
"craft/bookbinder": ("Bookbinder", "📖"),
"craft/brewery": ("Brewery", "🍺"),
"craft/builder": ("Builder", "🧱"),
"craft/carpenter": ("Carpenter", "🪚"),
"craft/caterer": ("Caterer", "🍱"),
"craft/cleaning": ("Cleaning Service", "🧹"),
"craft/confectionery": ("Confectioner", "🍬"),
"craft/distillery": ("Distillery", "🥃"),
"craft/dressmaker": ("Dressmaker", "👗"),
"craft/electrician": ("Electrician", ""),
"craft/electronics_repair": ("Electronics Repair", "🔌"),
"craft/floorer": ("Flooring Specialist", "🪵"),
"craft/gardener": ("Gardener", "🌱"),
"craft/glaziery": ("Glazier", "🪟"),
"craft/handicraft": ("Handicraft", "✂️"),
"craft/hvac": ("HVAC", "❄️"),
"craft/jeweller": ("Jeweller", "💎"),
"craft/joiner": ("Joiner", "🪚"),
"craft/key_cutter": ("Key Cutter", "🔑"),
"craft/locksmith": ("Locksmith", "🔐"),
"craft/metal_construction": ("Metal Fabrication", "🔩"),
"craft/painter": ("Painter & Decorator", "🖌️"),
"craft/photographer": ("Photographer", "📷"),
"craft/photographic_laboratory": ("Photo Lab", "🖼️"),
"craft/plumber": ("Plumber", "🔧"),
"craft/pottery": ("Pottery", "🏺"),
"craft/printer": ("Printer", "🖨️"),
"craft/roofer": ("Roofer", "🏠"),
"craft/sawmill": ("Sawmill", "🪵"),
"craft/scaffolder": ("Scaffolder", "🏗️"),
"craft/sculptor": ("Sculptor", "🗿"),
"craft/shoemaker": ("Shoemaker", "👞"),
"craft/signmaker": ("Sign Maker", "🪧"),
"craft/stonemason": ("Stonemason", "🪨"),
"craft/tailor": ("Tailor", "🧵"),
"craft/upholsterer": ("Upholsterer", "🛋️"),
"craft/watchmaker": ("Watchmaker", ""),
"craft/window_construction": ("Window Fitter", "🪟"),
"craft/winery": ("Winery", "🍷"),
"craft/yes": ("Craft Workshop", "🛠️"),
# healthcare
"healthcare/alternative": ("Alternative Medicine", "🌿"),
"healthcare/audiologist": ("Audiologist", "👂"),
"healthcare/centre": ("Health Centre", "🏥"),
"healthcare/clinic": ("Health Clinic", "🩺"),
"healthcare/counselling": ("Counselling", "🧠"),
"healthcare/dentist": ("Dental Practice", "🦷"),
"healthcare/doctor": ("GP Surgery", "👨‍⚕️"),
"healthcare/hospital": ("Hospital", "🏥"),
"healthcare/laboratory": ("Medical Lab", "🔬"),
"healthcare/optometrist": ("Optometrist", "👁️"),
"healthcare/pharmacy": ("Pharmacy", "💊"),
"healthcare/physiotherapist": ("Physiotherapist", "🏃"),
"healthcare/podiatrist": ("Podiatrist", "🦶"),
"healthcare/psychotherapist": ("Psychotherapist", "🧠"),
"healthcare/rehabilitation": ("Rehabilitation Centre", ""),
"healthcare/vaccination_centre": ("Vaccination Centre", "💉"),
"healthcare/yes": ("Healthcare Facility", "🏥"),
# leisure
"leisure/adult_gaming_centre": ("Adult Gaming Centre", "🎮"),
"leisure/amusement_arcade": ("Amusement Arcade", "🕹️"),
"leisure/bandstand": ("Bandstand", "🎺"),
"leisure/bathing_place": ("Bathing Spot", "🏖️"),
"leisure/bird_hide": ("Bird Hide", "🐦"),
"leisure/bowling_alley": ("Bowling Alley", "🎳"),
"leisure/common": ("Common Land", "🌳"),
"leisure/dance": ("Dance Venue", "💃"),
"leisure/dog_park": ("Dog Park", "🐕"),
"leisure/escape_game": ("Escape Room", "🔓"),
"leisure/fitness_centre": ("Gym", "🏋️"),
"leisure/fitness_station": ("Outdoor Gym", "💪"),
"leisure/garden": ("Garden", "🌷"),
"leisure/golf_course": ("Golf Course", ""),
"leisure/hackerspace": ("Hackerspace", "💻"),
"leisure/horse_riding": ("Horse Riding", "🐎"),
"leisure/indoor_play": ("Indoor Play Area", "🧒"),
"leisure/marina": ("Marina", ""),
"leisure/miniature_golf": ("Mini Golf", ""),
"leisure/nature_reserve": ("Nature Reserve", "🦔"),
"leisure/outdoor_seating": ("Outdoor Seating", "🪑"),
"leisure/park": ("Park", "🌳"),
"leisure/pitch": ("Sports Pitch", ""),
"leisure/playground": ("Playground", "🛝"),
"leisure/sauna": ("Sauna", "🧖"),
"leisure/slipway": ("Slipway", "🚤"),
"leisure/social_club": ("Social Club", "🍻"),
"leisure/sports_centre": ("Sports Centre", "🏟️"),
"leisure/sports_hall": ("Sports Hall", "🏀"),
"leisure/swimming_pool": ("Swimming Pool", "🏊"),
"leisure/tanning_salon": ("Tanning Salon", "☀️"),
"leisure/track": ("Running Track", "🏃"),
"leisure/trampoline_park": ("Trampoline Park", "🤸"),
"leisure/water_park": ("Water Park", "🌊"),
"leisure/wildlife_hide": ("Wildlife Hide", "🦌"),
"leisure/yes": ("Leisure Facility", "🎉"),
# office
"office/accountant": ("Accountant", "🧮"),
"office/advertising_agency": ("Advertising Agency", "📢"),
"office/architect": ("Architect", "📐"),
"office/association": ("Association", "🏛️"),
"office/charity": ("Charity", "❤️"),
"office/construction_company": ("Construction Company", "🏗️"),
"office/consulting": ("Consulting Firm", "📊"),
"office/courier": ("Courier Service", "📦"),
"office/coworking": ("Co-working Space", "💻"),
"office/design": ("Design Studio", "🎨"),
"office/diplomatic": ("Diplomatic Office", "🏛️"),
"office/educational_institution": ("Education Office", "🎓"),
"office/employment_agency": ("Employment Agency", "💼"),
"office/energy_supplier": ("Energy Supplier", ""),
"office/engineer": ("Engineering Firm", "⚙️"),
"office/estate_agent": ("Estate Agent", "🏠"),
"office/financial": ("Financial Services", "💰"),
"office/financial_advisor": ("Financial Advisor", "📈"),
"office/foundation": ("Foundation", "🏛️"),
"office/government": ("Government Office", "🏛️"),
"office/graphic_design": ("Graphic Design", "🖌️"),
"office/healthcare": ("Healthcare Office", "🏥"),
"office/home_care": ("Home Care Service", "🏠"),
"office/insurance": ("Insurance", "🛡️"),
"office/interior_design": ("Interior Design", "🛋️"),
"office/it": ("IT Company", "💻"),
"office/lawyer": ("Lawyer", "⚖️"),
"office/logistics": ("Logistics", "🚚"),
"office/marketing": ("Marketing Agency", "📣"),
"office/mortgage": ("Mortgage Broker", "🏦"),
"office/moving_company": ("Moving Company", "📦"),
"office/newspaper": ("Newspaper Office", "📰"),
"office/ngo": ("NGO", "🌍"),
"office/notary": ("Notary", "📜"),
"office/political_party": ("Political Party", "🗳️"),
"office/politician": ("Politician Office", "🏛️"),
"office/property_management": ("Property Management", "🏘️"),
"office/recruitment": ("Recruitment Agency", "👥"),
"office/religion": ("Religious Office", "✝️"),
"office/research": ("Research Office", "🔬"),
"office/security": ("Security Company", "🔒"),
"office/solicitor": ("Solicitor", "⚖️"),
"office/surveyor": ("Surveyor", "📏"),
"office/tax_advisor": ("Tax Advisor", "🧾"),
"office/taxi": ("Taxi Office", "🚕"),
"office/telecommunication": ("Telecoms Office", "📡"),
"office/therapist": ("Therapist", "🧠"),
"office/travel_agent": ("Travel Agent", "✈️"),
"office/union": ("Trade Union", ""),
"office/university": ("University Office", "🎓"),
"office/vacant": ("Vacant Office", "🏚️"),
"office/web_design": ("Web Design", "🌐"),
# public_transport
"public_transport/entrance": ("Transport Entrance", "🚪"),
"public_transport/platform": ("Platform", "🚉"),
"public_transport/station": ("Station", "🚉"),
"public_transport/stop_position": ("Stop", "🚏"),
# shop
"shop/accessories": ("Accessories Shop", "👜"),
"shop/agrarian": ("Farm Supply Shop", "🌾"),
"shop/alcohol": ("Off-Licence", "🍷"),
"shop/antiques": ("Antiques Shop", "🏺"),
"shop/appliance": ("Appliance Shop", "🔌"),
"shop/art": ("Art Shop", "🎨"),
"shop/baby_goods": ("Baby Shop", "🍼"),
"shop/bag": ("Bag Shop", "👜"),
"shop/bakery": ("Bakery", "🥐"),
"shop/bathroom": ("Bathroom Shop", "🛁"),
"shop/bathroom_furnishing": ("Bathroom Furnishings", "🚿"),
"shop/beauty": ("Beauty Shop", "💄"),
"shop/bed": ("Bed Shop", "🛏️"),
"shop/beverages": ("Drinks Shop", "🥤"),
"shop/bicycle": ("Bike Shop", "🚲"),
"shop/boat": ("Boat Shop", ""),
"shop/bookmaker": ("Bookmaker", "🏇"),
"shop/books": ("Bookshop", "📚"),
"shop/boutique": ("Boutique", "👗"),
"shop/building_materials": ("Building Materials", "🧱"),
"shop/butcher": ("Butcher", "🥩"),
"shop/camera": ("Camera Shop", "📷"),
"shop/candles": ("Candle Shop", "🕯️"),
"shop/car": ("Car Dealership", "🚗"),
"shop/car;car_repair": ("Car Sales & Repair", "🚗"),
"shop/car_parts": ("Car Parts", "🔩"),
"shop/car_repair": ("Car Repair", "🔧"),
"shop/caravan": ("Caravan Dealer", "🚐"),
"shop/carpet": ("Carpet Shop", "🧶"),
"shop/catalogue": ("Catalogue Shop", "📋"),
"shop/charity": ("Charity Shop", "❤️"),
"shop/cheese": ("Cheese Shop", "🧀"),
"shop/chemist": ("Chemist", "🧪"),
"shop/chocolate": ("Chocolate Shop", "🍫"),
"shop/clothes": ("Clothes Shop", "👕"),
"shop/coffee": ("Coffee Shop", ""),
"shop/collector": ("Collector Shop", "🏆"),
"shop/computer": ("Computer Shop", "🖥️"),
"shop/confectionery": ("Sweet Shop", "🍬"),
"shop/convenience": ("Convenience Store", "🏪"),
"shop/copyshop": ("Copy Shop", "🖨️"),
"shop/cosmetics": ("Cosmetics Shop", "💅"),
"shop/country_store": ("Country Store", "🏡"),
"shop/craft": ("Craft Shop", "✂️"),
"shop/curtain": ("Curtain Shop", "🪟"),
"shop/dairy": ("Dairy Shop", "🥛"),
"shop/deli": ("Delicatessen", "🧆"),
"shop/department_store": ("Department Store", "🏬"),
"shop/discount": ("Discount Store", "💲"),
"shop/doityourself": ("DIY Store", "🔨"),
"shop/doors": ("Door Shop", "🚪"),
"shop/dry_cleaning": ("Dry Cleaner", "👔"),
"shop/e-cigarette": ("Vape Shop", "💨"),
"shop/electrical": ("Electrical Shop", ""),
"shop/electronics": ("Electronics Shop", "📱"),
"shop/erotic": ("Adult Shop", "🔞"),
"shop/esoteric": ("Esoteric Shop", "🔮"),
"shop/estate_agent": ("Estate Agent", "🏠"),
"shop/fabric": ("Fabric Shop", "🧵"),
"shop/fan": ("Fan Shop", "🏅"),
"shop/farm": ("Farm Shop", "🥕"),
"shop/fashion_accessories": ("Fashion Accessories", "👒"),
"shop/fireplace": ("Fireplace Shop", "🔥"),
"shop/fishing": ("Fishing Shop", "🎣"),
"shop/flooring": ("Flooring Shop", "🪵"),
"shop/florist": ("Florist", "💐"),
"shop/food": ("Food Shop", "🍞"),
"shop/frame": ("Framing Shop", "🖼️"),
"shop/frozen_food": ("Frozen Food Shop", "🧊"),
"shop/fuel": ("Fuel Shop", ""),
"shop/funeral_directors": ("Funeral Director", "⚰️"),
"shop/furniture": ("Furniture Shop", "🪑"),
"shop/games": ("Games Shop", "🎮"),
"shop/garden_centre": ("Garden Centre", "🌻"),
"shop/gas": ("Gas Shop", "🔥"),
"shop/general": ("General Store", "🏪"),
"shop/gift": ("Gift Shop", "🎁"),
"shop/glaziery": ("Glazier", "🪟"),
"shop/greengrocer": ("Greengrocer", "🥬"),
"shop/grocery": ("Grocery Shop", "🛒"),
"shop/haberdashery": ("Haberdashery", "🧵"),
"shop/hairdresser": ("Hairdresser", "💇"),
"shop/hairdresser_supply": ("Hairdresser Supply", "💇"),
"shop/hardware": ("Hardware Shop", "🔩"),
"shop/health": ("Health Shop", "🌿"),
"shop/health_food": ("Health Food Shop", "🥗"),
"shop/hearing_aids": ("Hearing Aid Shop", "👂"),
"shop/herbalist": ("Herbalist", "🌿"),
"shop/hifi": ("Hi-Fi Shop", "🔊"),
"shop/household": ("Household Shop", "🏠"),
"shop/household_linen": ("Linen Shop", "🛏️"),
"shop/houseware": ("Houseware Shop", "🍳"),
"shop/ice_cream": ("Ice Cream Shop", "🍦"),
"shop/interior_decoration": ("Interior Decoration", "🖼️"),
"shop/jewelry": ("Jewellery Shop", "💍"),
"shop/kiosk": ("Kiosk", "🏪"),
"shop/kitchen": ("Kitchen Shop", "🍳"),
"shop/laundry": ("Laundry", "🧺"),
"shop/leather": ("Leather Shop", "🧳"),
"shop/lighting": ("Lighting Shop", "💡"),
"shop/locksmith": ("Locksmith", "🔐"),
"shop/mall": ("Shopping Centre", "🏬"),
"shop/massage": ("Massage Parlour", "💆"),
"shop/medical_supply": ("Medical Supply", "🩺"),
"shop/military_surplus": ("Military Surplus", "🎖️"),
"shop/mobile_phone": ("Mobile Phone Shop", "📱"),
"shop/mobile_phone_accessories": ("Phone Accessories", "📱"),
"shop/mobility": ("Mobility Shop", ""),
"shop/mobility_scooter": ("Mobility Scooter Shop", "🦽"),
"shop/model": ("Model Shop", "✈️"),
"shop/money_lender": ("Money Lender", "💰"),
"shop/motorcycle": ("Motorcycle Shop", "🏍️"),
"shop/motorcycle_repair": ("Motorcycle Repair", "🔧"),
"shop/music": ("Music Shop", "🎵"),
"shop/musical_instrument": ("Musical Instrument Shop", "🎸"),
"shop/newsagent": ("Newsagent", "📰"),
"shop/nutrition_supplements": ("Nutrition Shop", "💪"),
"shop/optician": ("Optician", "👓"),
"shop/outdoor": ("Outdoor Shop", "🏕️"),
"shop/outpost": ("Outpost", "📦"),
"shop/paint": ("Paint Shop", "🎨"),
"shop/party": ("Party Shop", "🎈"),
"shop/pastry": ("Pastry Shop", "🥐"),
"shop/pawnbroker": ("Pawnbroker", "💰"),
"shop/perfumery": ("Perfumery", "🌸"),
"shop/pet": ("Pet Shop", "🐾"),
"shop/pet_grooming": ("Pet Grooming", "🐩"),
"shop/photo": ("Photo Shop", "📸"),
"shop/piercing": ("Piercing Studio", "💎"),
"shop/plant_hire": ("Plant Hire", "🚜"),
"shop/pottery": ("Pottery Shop", "🏺"),
"shop/printer_ink": ("Ink & Toner Shop", "🖨️"),
"shop/printing": ("Print Shop", "🖨️"),
"shop/psychic": ("Psychic", "🔮"),
"shop/pyrotechnics": ("Fireworks Shop", "🎆"),
"shop/religion": ("Religious Shop", "✝️"),
"shop/rental": ("Rental Shop", "🔑"),
"shop/repair": ("Repair Shop", "🔧"),
"shop/scuba_diving": ("Scuba Diving Shop", "🤿"),
"shop/seafood": ("Fishmonger", "🐟"),
"shop/second_hand": ("Second-Hand Shop", "♻️"),
"shop/security": ("Security Shop", "🔒"),
"shop/sewing": ("Sewing Shop", "🪡"),
"shop/shoe_repair": ("Shoe Repair", "👞"),
"shop/shoes": ("Shoe Shop", "👟"),
"shop/sports": ("Sports Shop", ""),
"shop/stationery": ("Stationery Shop", "✏️"),
"shop/storage_rental": ("Self Storage", "📦"),
"shop/supermarket": ("Supermarket", "🛒"),
"shop/swimming_pool": ("Pool Supplies", "🏊"),
"shop/tailor": ("Tailor", "🧵"),
"shop/tattoo": ("Tattoo Studio", "🖋️"),
"shop/taxi": ("Taxi Booking", "🚕"),
"shop/tea": ("Tea Shop", "🫖"),
"shop/telecommunication": ("Telecoms Shop", "📡"),
"shop/ticket": ("Ticket Office", "🎫"),
"shop/tiles": ("Tile Shop", "🔲"),
"shop/tobacco": ("Tobacconist", "🚬"),
"shop/tool_hire": ("Tool Hire", "🧰"),
"shop/toys": ("Toy Shop", "🧸"),
"shop/trade": ("Trade Supplier", "🏭"),
"shop/travel_agency": ("Travel Agency", "✈️"),
"shop/trophy": ("Trophy Shop", "🏆"),
"shop/tyres": ("Tyre Shop", "🛞"),
"shop/vacant": ("Vacant Shop", "🏚️"),
"shop/variety_store": ("Variety Store", "🏪"),
"shop/video": ("Video Shop", "📀"),
"shop/video_games": ("Video Game Shop", "🎮"),
"shop/watches": ("Watch Shop", ""),
"shop/water_sports": ("Water Sports Shop", "🏄"),
"shop/weapons": ("Weapons Shop", "🗡️"),
"shop/wedding": ("Wedding Shop", "💒"),
"shop/wholesale": ("Wholesaler", "📦"),
"shop/wigs": ("Wig Shop", "💇"),
"shop/window_blind": ("Blinds Shop", "🪟"),
"shop/windows": ("Window Shop", "🪟"),
"shop/wine": ("Wine Shop", "🍷"),
"shop/wool": ("Wool Shop", "🧶"),
"shop/yes": ("Shop", "🛍️"),
# tourism
"tourism/artwork": ("Public Artwork", "🎨"),
"tourism/attraction": ("Tourist Attraction", "📸"),
"tourism/camp_site": ("Campsite", ""),
"tourism/caravan_site": ("Caravan Site", "🚐"),
"tourism/chalet": ("Chalet", "🏔️"),
"tourism/gallery": ("Gallery", "🖼️"),
"tourism/guest_house": ("Guest House", "🏡"),
"tourism/hostel": ("Hostel", "🛏️"),
"tourism/hotel": ("Hotel", "🏨"),
"tourism/motel": ("Motel", "🏨"),
"tourism/museum": ("Museum", "🏛️"),
"tourism/picnic_site": ("Picnic Site", "🧺"),
"tourism/preserved_railway": ("Heritage Railway", "🚂"),
"tourism/theme_park": ("Theme Park", "🎢"),
"tourism/viewpoint": ("Viewpoint", "🔭"),
"tourism/zoo": ("Zoo", "🦁"),
}
def transform(input_path: Path) -> pl.LazyFrame:
lf = pl.scan_parquet(input_path)
# Get all unique categories present in the data
all_categories = lf.select("category").unique().collect().to_series().to_list()
# Verify every non-dropped category has a mapping
unmapped = []
for cat in all_categories:
if cat not in DROP_CATEGORIES and cat not in CATEGORY_MAP:
unmapped.append(cat)
if unmapped:
raise ValueError(
f"Categories missing from CATEGORY_MAP: {sorted(unmapped)}"
)
# Verify every CATEGORY_MAP key actually exists in the data (catch typos)
mapped_but_absent = []
all_set = set(all_categories)
for cat in CATEGORY_MAP:
if cat not in all_set:
mapped_but_absent.append(cat)
if mapped_but_absent:
raise ValueError(
f"CATEGORY_MAP contains categories not in data: {sorted(mapped_but_absent)}"
)
# Drop unwanted categories
lf = lf.filter(~pl.col("category").is_in(list(DROP_CATEGORIES)))
# Build name and emoji lookup expressions
name_mapping = {k: v[0] for k, v in CATEGORY_MAP.items()}
emoji_mapping = {k: v[1] for k, v in CATEGORY_MAP.items()}
# Check no friendly names are missing (defensive)
missing_names = [k for k, v in CATEGORY_MAP.items() if not v[0]]
if missing_names:
raise ValueError(f"Empty friendly names for: {missing_names}")
missing_emojis = [k for k, v in CATEGORY_MAP.items() if not v[1]]
if missing_emojis:
raise ValueError(f"Empty emojis for: {missing_emojis}")
lf = lf.with_columns(
pl.col("category").replace_strict(name_mapping).alias("category"),
pl.col("category").replace_strict(emoji_mapping).alias("emoji"),
)
return lf
def main():
parser = argparse.ArgumentParser(description="Transform raw POIs to filtered version with friendly names")
parser.add_argument("--input", type=Path, required=True, help="Raw POIs parquet file")
parser.add_argument("--output", type=Path, required=True, help="Output filtered POIs parquet file")
args = parser.parse_args()
df = transform(args.input).collect()
df.write_parquet(args.output)
size_mb = args.output.stat().st_size / (1024 * 1024)
print(f"Wrote {args.output} ({size_mb:.1f} MB, {len(df):,} POIs)")
print(f"\nCategories ({df['category'].n_unique()}):")
counts = df.group_by("category", "emoji").len().sort("len", descending=True)
for row in counts.iter_rows(named=True):
print(f" {row['emoji']} {row['category']}: {row['len']:,}")
if __name__ == "__main__":
main()

View file

@ -6,7 +6,7 @@ POSTCODE = "E14 2DG"
# Price paid: unique addresses for this postcode
pp = (
pl.scan_parquet("data_sources/pp-complete.parquet")
pl.scan_parquet("data/price-paid-complete.parquet")
.filter(pl.col("postcode") == POSTCODE)
.select("paon", "saon", "street", "postcode")
.unique()
@ -22,7 +22,7 @@ pp = (
# EPC: latest inspection per address for this postcode
epc = (
pl.scan_csv("data_sources/epc/certificates.csv")
pl.scan_csv("data/epc/certificates.csv")
.select("ADDRESS", "POSTCODE", "INSPECTION_DATE")
.filter(pl.col("POSTCODE").str.strip_chars() == POSTCODE)
.sort("INSPECTION_DATE", descending=True)

View file

@ -1,143 +0,0 @@
"""Build a wide property dataframe and H3 aggregates from epc_pp output."""
import polars as pl
import h3
from pipeline.config import AGGREGATES_DIR, H3_RESOLUTIONS, DATA_DIR, PROCESSED_DIR
def _build_wide() -> pl.DataFrame:
"""Build the wide dataframe by joining epc_pp with all auxiliary data."""
print("Loading epc_pp...")
wide = pl.read_parquet(PROCESSED_DIR / "epc_pp.parquet")
print(f" {wide.shape[0]:,} rows")
# GPS coordinates + LSOA from ArcGIS
print("Joining GPS coordinates...")
arcgis = pl.read_parquet(DATA_DIR / "arcgis_data.parquet").select(
pl.col("pcds").alias("postcode"),
"lat",
pl.col("long").alias("lon"),
"lsoa21",
)
wide = wide.join(arcgis, on="postcode", how="inner")
print(f" {wide.shape[0]:,} rows after GPS join")
# Journey times (optional)
journey_path = PROCESSED_DIR / "journey_times_bank_checkpoint.parquet"
if journey_path.exists():
print("Joining journey times...")
journey_times = pl.read_parquet(journey_path).select(
"postcode",
"public_transport_easy_minutes",
"public_transport_quick_minutes",
"cycling_minutes",
)
wide = wide.join(journey_times, on="postcode", how="left")
# Index of Deprivation
iod_path = DATA_DIR / "IoD2025_Scores.parquet"
if iod_path.exists():
print("Joining IoD scores...")
iod = pl.read_parquet(iod_path).drop(
"LSOA name (2021)",
"Local Authority District code (2024)",
"Local Authority District name (2024)",
)
# Rename IoD columns to clean snake_case
iod = iod.rename(_IOD_RENAMES)
wide = wide.join(
iod, left_on="lsoa21", right_on="lsoa_code", how="left"
)
return wide
_IOD_RENAMES = {
"LSOA code (2021)": "lsoa_code",
"Index of Multiple Deprivation (IMD) Score": "imd_score",
"Income Score (rate)": "income_score",
"Employment Score (rate)": "employment_score",
"Education, Skills and Training Score": "education_score",
"Health Deprivation and Disability Score": "health_score",
"Crime Score": "crime_score",
"Barriers to Housing and Services Score": "housing_barriers_score",
"Living Environment Score": "living_environment_score",
"Income Deprivation Affecting Children Index (IDACI) Score (rate)": "idaci_score",
"Income Deprivation Affecting Older People (IDAOPI) Score (rate)": "idaopi_score",
"Children and Young People Sub-domain Score": "children_young_people_score",
"Adult Skills Sub-domain Score": "adult_skills_score",
"Geographical Barriers Sub-domain Score": "geographical_barriers_score",
"Wider Barriers Sub-domain Score": "wider_barriers_score",
"Indoors Sub-domain Score": "indoors_score",
"Outdoors Sub-domain Score": "outdoors_score",
}
def _add_h3_indices(df: pl.DataFrame) -> pl.DataFrame:
"""Compute H3 indices from lat/lon for all configured resolutions."""
print("Computing H3 indices...")
# Compute per unique postcode for efficiency, then join back
postcodes = df.select("postcode", "lat", "lon").unique(subset=["postcode"])
for res in H3_RESOLUTIONS:
col_name = f"h3_res{res}"
postcodes = postcodes.with_columns(
pl.struct(["lat", "lon"])
.map_elements(
lambda x, r=res: h3.latlng_to_cell(x["lat"], x["lon"], r),
return_dtype=pl.Utf8,
)
.alias(col_name)
)
print(f" res{res}: {postcodes[col_name].n_unique():,} unique cells")
h3_cols = [f"h3_res{res}" for res in H3_RESOLUTIONS]
return df.join(
postcodes.select("postcode", *h3_cols), on="postcode", how="left"
)
def _aggregate_to_h3(df: pl.DataFrame) -> None:
"""Aggregate min/max of every numeric feature per H3 cell at each resolution."""
AGGREGATES_DIR.mkdir(parents=True, exist_ok=True)
exclude = {"lat", "lon"}
numeric_cols = [
col
for col, dtype in zip(df.columns, df.dtypes)
if dtype.is_numeric() and not col.startswith("h3_res") and col not in exclude
]
agg_exprs = [pl.len().alias("count")]
for col in numeric_cols:
agg_exprs.append(pl.col(col).min().alias(f"min_{col}"))
agg_exprs.append(pl.col(col).max().alias(f"max_{col}"))
print("Aggregating to H3 cells...")
for res in H3_RESOLUTIONS:
h3_col = f"h3_res{res}"
result = df.group_by(h3_col).agg(agg_exprs).rename({h3_col: "h3"})
path = AGGREGATES_DIR / f"res{res}.parquet"
result.write_parquet(path)
size_mb = path.stat().st_size / (1024 * 1024)
print(f" {path.name}: {result.shape[0]:,} cells ({size_mb:.1f} MB)")
def run():
"""Run the full wide pipeline: build wide df, compute H3, aggregate."""
wide = _build_wide()
wide_path = PROCESSED_DIR / "wide.parquet"
wide.write_parquet(wide_path)
size_mb = wide_path.stat().st_size / (1024 * 1024)
print(f"Wrote {wide_path} ({size_mb:.1f} MB)")
wide = _add_h3_indices(wide)
_aggregate_to_h3(wide)
print("Done.")
if __name__ == "__main__":
run()