Data catalog

# Urban Growth Center (UGC) data

Context for working with UGC data from Development Data Lab: city-level datasets for urban areas worldwide, published as independently versioned modules.

Catalog: https://cities.devdatalab.org/data
Documentation: https://cities.devdatalab.org/docs/datasets

## Getting the files
- Each link is a .zip holding the data file and a README with the full schema, license and disclaimer. Download and unzip before reading.
- Links redirect to the file host. Follow redirects (e.g. `curl -L -O <url>`).

## Core concepts
- `id` identifies a boundary polygon and is unique across PARENT, ADM1 and ADM2. Join tables on `id` alone.
- `boundary_level` is PARENT (the whole city), ADM1 or ADM2 (the city split by administrative unit). Use it to select one level, not as a join key. Filter to a single level before aggregating, or values will be double counted.
- ADM1/ADM2 split polygons carry `parent_id`. Splits may not cover the whole parent, so do not assume split values sum to the parent total.
- Tables with a `version` column record the boundary version they were built on. Modules are versioned independently, so check that boundary versions agree before joining across modules.
- Null means unavailable, not zero.

## Datasets

### DDL Global Urban Boundaries (`ua-boundaries` v0.3.0)
City polygons with stable identifiers, published globally.
- Levels: PARENT, ADM1, ADM2
- Unique key: id
- Format: GeoPackage (.gpkg) in a .zip
- Coverage: Every parent city with a population of at least 50,000.
- Columns: id, name, population (people), alpha3, area (m²), geometry
- Files:
  - Parent boundaries: https://cities.devdatalab.org/downloads/ua-boundaries/v0.3.0/ddl-urban-boundaries-global-v0.3.0.zip
  - ADM1 split: https://cities.devdatalab.org/downloads/ua-boundaries/v0.3.0/ddl-urban-boundaries-adm1-global-v0.3.0.zip
  - ADM2 split: https://cities.devdatalab.org/downloads/ua-boundaries/v0.3.0/ddl-urban-boundaries-adm2-global-v0.3.0.zip
- Notes:
  - The parent city is the reference geography for the city profile. Its identifier connects the published tables.
  - ADM1 and ADM2 files contain released administrative intersections of qualifying parent polygons. Published splits may not cover the whole parent; do not assume their measures sum to parent totals. Use the split identifier for split-level measures and parent_id to relate a split to its parent.
  - GHSL and GADM map overlays provide comparison geographies. They do not change the boundaries used to calculate the profile measures.
  - area is computed on the geography type, in square metres, not in the layer's own projected units.
  - Hong Kong publishes under CHN, because the boundaries carry it that way.
- Source: DDL Urban Growth Center (https://www.devdatalab.org)
- Docs: https://cities.devdatalab.org/docs/datasets/ua-boundaries

### Historical Population (`population-historical` v0.3.0)
Population totals derived from GHSL population raster data, by year.
- Levels: PARENT, ADM1, ADM2
- Unique key: id, version, year
- Years: 1975–2025
- Format: Parquet in a .zip
- Coverage: Every DDL boundary with a population raster overlap, at parent, ADM1 and ADM2 split levels.
- Columns: id, version, boundary_level, year, pop_sum (People), pixel_count
- Files:
  - Full table: https://cities.devdatalab.org/downloads/population-historical/v0.3.0/ddl-population-historical-v0.3.0.zip
- Notes:
  - The population history follows the same released boundary through time. Growth describes population change inside that footprint; it does not measure how the urban boundary expanded.
  - Years between the five-year GHS-POP epochs are linearly interpolated by UGC, and are not measured directly in the upstream dataset.
- Source: GHS-POP R2023A — GHS population grid multitemporal (1975–2030) (https://human-settlement.emergency.copernicus.eu/ghs_pop2023.php)
- Docs: https://cities.devdatalab.org/docs/datasets/population-historical

### Urban PM2.5, monthly (`pm25-monthly` v0.3.0)
Monthly mean surface PM2.5 concentrations derived from ACAG gridded data.
- Levels: PARENT, ADM1, ADM2
- Unique key: id, version, pm25_year, pm25_month
- Years: 1998–2024
- Format: Parquet in a .zip
- Coverage: Every DDL boundary that overlaps the upstream grid, at parent, ADM1 and ADM2 level.
- Columns: id, version, boundary_level, pm25_year, pm25_month, pm25_version, pm25_mean (µg/m³), pm25_cell_count
- Files:
  - Full table: https://cities.devdatalab.org/downloads/pm25-monthly/v0.3.0/ddl-pm25-monthly-v0.3.0.zip
- Notes:
  - Partial grid cells are area-weighted when calculating the mean surface PM2.5 for the polygon.
- Source: Washington University in St. Louis, Atmospheric Composition Analysis Group — surface PM2.5 (https://sites.wustl.edu/acag/datasets/surface-pm2-5/)
- Docs: https://cities.devdatalab.org/docs/datasets/pm25-monthly

### Urban Congestion Indices (`congestion` v0.3.0)
Speed and congestion indices derived from Google Maps.
- Levels: PARENT
- Unique key: id, version
- Years: 2019
- Format: Parquet in a .zip
- Coverage: Not all cities have values. Authors' source cities are mapped onto cities and aggregated as appropriate to produce one value per city.
- Columns: id, version, traffic_year, congestion_index, speed_index, uncongested_speed_index, speed_index_sd, mean_speed_kmh (km/h)
- Files:
  - Full table: https://cities.devdatalab.org/downloads/congestion/v0.3.0/ddl-congestion-v0.3.0.zip
- Notes:
  - Data is only available for a subset of cities
  - Traffic data comes from measurements taken between June and November 2019 on Google maps from simulated trips
  - Several source-city polygons may contribute to one city. Look at `speed_index_sd` for spread across contributing cities
  - uncongested_speed_index = speed_index + congestion_index, exactly. It is a decomposition, not an independent measurement.
- Source: Akbar et al. (2023) — The Fast, the Slow, and the Congested: Urban Transportation in Rich and Poor Countries (https://www.nber.org/papers/w31642)
- Docs: https://cities.devdatalab.org/docs/datasets/congestion

### GHSL Built-up Zonal Statistics (`ghsl-builtup` v0.3.0)
Built-up surface, volume, non-residential use, and height derived from GHSL raster data.
- Levels: PARENT, ADM1, ADM2
- Unique key: id, version, ghsl_year
- Years: 1975, 1980, 1985, 1990, 1995, 2000, 2005, 2010, 2015, 2018, 2020, 2025
- Format: Parquet in a .zip
- Coverage: Includes released parent cities and their ADM1 and ADM2 splits when the source raster metric is available.
- Columns: id, version, boundary_level, ghsl_release, ghsl_year, built_up_surface_sum_m2 (m²), built_up_surface_avg_m2 (m²), built_up_volume_sum_m3 (m³), built_up_volume_avg_m3 (m³), nonres_built_up_surface_sum_m2 (m²), nonres_built_up_surface_avg_m2 (m²), nonres_built_up_volume_sum_m3 (m³), nonres_built_up_volume_avg_m3 (m³), gross_built_up_height_avg_m (m), net_built_up_height_avg_m (m), built_up_surface_valid_coverage_share, built_up_volume_valid_coverage_share, nonres_built_up_surface_valid_coverage_share, nonres_built_up_volume_valid_coverage_share, gross_built_up_height_valid_coverage_share, net_built_up_height_valid_coverage_share
- Files:
  - Full table: https://cities.devdatalab.org/downloads/ghsl-builtup/v0.3.0/ddl-ghsl-builtup-v0.3.0.zip
- Notes:
  - `ghsl_year` is 1975/1980/1985/1990/1995/2000/2005/2010/2015/2020/2025 for Built-up surface and volume data; gross and net height come from GHS-BUILT-H, which is from 2018.
  - Selecting on `ghsl_year` is therefore necessary to select the appropriate raster product. 2018 will not have any built-up surface or volume data while the other years will not have any height data.
  - Filter boundary_level before aggregation so parent and administrative zones are not double counted.
  - The built-up surface and volume metrics for 2025 are projections from the 2020 data.
  - The building volume measure is mechanically constructed by multiplying the built-up surface by the net built-up height in 2018. Therefore the ratio of built up volume to surface is the same across all years.
  - Net building height is calculated as built-up volume over valid built-up surface area. Gross building height is calculated as built-up volume over the entire surface area.
  - Keep every metric with its corresponding validity-coverage share; null metrics are unavailable, not zero. A share of 0 means the raster was absent or no valid pixel fell inside the zone.
- Source: European Commission, Joint Research Centre — Global Human Settlement Layer, GHS-BUILT-S, GHS-BUILT-V and GHS-BUILT-H R2023A (https://human-settlement.emergency.copernicus.eu/p2023Release.php)
- Docs: https://cities.devdatalab.org/docs/datasets/ghsl-builtup

### World Bank Subnational Doing Business (`doing-business` v0.3.0)
Business regulation scores from the World Bank’s Subnational Doing Business surveys.
- Levels: PARENT
- Unique key: id, version, alpha3, source_city, data_as_of
- Years: 2009–2022
- Format: Parquet in a .zip
- Coverage: Only parent cities matched from a World Bank source city are included. Several source-city observations may map to one city.
- Columns: id, version, alpha3, source_country_name, source_city, source_region, data_as_of, starting_business_overall_score, construction_permit_overall_score_20, construction_permit_overall_score_15, getting_electricity_overall_score_20, getting_electricity_overall_score_15, registering_property_overall_score_20, registering_property_overall_score_15, paying_taxes_overall_score_20, paying_taxes_overall_score_16, enforcing_contracts_overall_score_20, enforcing_contracts_overall_score_15
- Files:
  - Full table: https://cities.devdatalab.org/downloads/doing-business/v0.3.0/ddl-doing-business-v0.3.0.zip
- Notes:
  - Some of the measures have suffixes such as _20, _16 or _15. These correspond to different iterations of the World Bank Score calculation methodology and thus should not be combined into one.
  - Several source cities may map to one city. The recommendation is to aggregate them at the city level as per the data need.
  - If a field is null, it means the score is unavailable for that observation in the source data.
- Source: World Bank — Subnational Doing Business (https://datacatalog.worldbank.org/search/dataset/0038564/doing-business)
- Docs: https://cities.devdatalab.org/docs/datasets/doing-business

## License and citation
- License: Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
- Cite: Development Data Lab. (2026). Urban Growth Center. https://cities.devdatalab.org/
- Also cite the original source listed under each dataset.