Weighing 600 Million Buildings Reveals Global Urban Material Inequality and Efficiency Paths

This dataset contains the city-level building material stock estimates used in the manuscript entitled "Weighing 600 Million Buildings Reveals Global Urban Material Inequality and Efficiency Paths." The data were generated using geospatial and statistical methods described in the manuscript. Please refer to the publication for methodological details and guidance on data interpretation and reuse. This v3 record contains the original deposited data files and a concise guide to access and interpretation. The complete README, including detailed instructions for data access, file organization, software use, field definitions, interpretation, limitations, licensing, and citation, is provided with version 4 of this dataset. Use the Concept DOI above to access the dataset version history and locate version 4. CONTACT Dataset creator: Jinchao SongEmail: jinchaos@umich.eduORCID: https://orcid.org/0000-0002-3448-124X For questions about the data or their interpretation, please contact the dataset creator at the email address above. OVERVIEW This repository contains individual-building-level material stock estimates for buildings located within the study cities. The data support the manuscript “Weighing 600 Million Buildings Reveals Global Urban Material Inequality and Efficiency Paths” and the study’s assessment of global urban material inequality and material-efficient urbanization pathways. The data are distributed as internal components of an Esri File Geodatabase. Each point record is intended to represent an individual building; however, some records may represent merged adjacent buildings owing to preprocessing. The material fields report modeled building material stocks in kilograms (kg). REPOSITORY INFORMATION Zenodo record: https://zenodo.org/records/19376248Version-specific DOI: https://doi.org/10.5281/zenodo.19376248Concept DOI (all versions): https://doi.org/10.5281/zenodo.15809010Dataset version: v3Access: OpenLicence: Creative Commons Attribution 4.0 International (CC BY 4.0)Licence URL: https://creativecommons.org/licenses/by/4.0/Number of deposited files: 59Total deposited size: approximately 20.1 GB Use the version-specific DOI when referring to or reproducing the exact v3 files. The Concept DOI may be used when referring to the dataset across all versions. FILE STRUCTURE AND DOWNLOAD The 59 files displayed separately on Zenodo are not 59 independent datasets. They are internal File Geodatabase components and must be downloaded together. Do not rename, edit, or interpret individual files whose names begin with “a000”. To reconstruct the deposited dataset: 1. Download all 59 files from the Zenodo record.2. Create a new folder with a name ending in “.gdb”, for example “dataset.gdb”.3. Place all 59 files directly inside that folder.4. Keep every downloaded filename unchanged.5. Open the resulting folder as an Esri File Geodatabase. Allow at least 30 GB of free disk space for downloading and reconstructing the dataset, plus additional space for temporary files or exported derivatives. For reliable processing, use a local solid-state drive rather than a synchronized cloud folder or unstable network drive. The two largest deposited files are: • a00000013.gdbtable — approximately 19.19 GB; MD5: b53550c6ca7eadbe8e72586ae6da1c18• a00000013.gdbtablx — approximately 0.92 GB; MD5: 52fff94acad28dcf8bcbf19d55b7cb3a Zenodo provides an MD5 checksum for every deposited file. If a download is interrupted or the database cannot be opened, compare the downloaded checksums with those displayed on the Zenodo record. OPENING THE DATA ArcGIS Pro: 1. Open the Catalog pane.2. Under Databases, choose Add Database.3. Select the reconstructed folder ending in “.gdb”.4. Expand the database and add the feature class to a map.5. Inspect the attribute table, coordinate reference system, and field definitions before analysis. QGIS: 1. Open Data Source Manager and select Vector.2. Set Source type to Directory.3. Select OpenFileGDB.4. Browse to the reconstructed folder ending in “.gdb”.5. Select the available layer. The complete dataset is very large. List the available layers and inspect a small sample before attempting to load all records. Use spatial or attribute filters and select only the fields needed for the analysis. ATTRIBUTE DICTIONARY OBJECTID — Integer. File Geodatabase-managed row identifier. It should not be treated as a persistent building identifier across exports or dataset versions. Shape — Point Z geometry. Building point location. The geometry is Z-enabled, but users should not assume that the Z coordinate represents building height without confirmation from the methods or metadata. Steel — Numeric, kg. Estimated steel stock. Wood — Numeric, kg. Estimated wood stock. Cement — Numeric, kg. Estimated cement stock. Stone — Numeric, kg. Estimated stone stock. Sand — Numeric, kg. Estimated sand stock. Brick — Numeric, kg. Estimated brick stock. Mortar_plaster — Numeric, kg. Estimated mortar and plaster stock. Glass — Numeric, kg. Estimated glass stock. Bitumen — Numeric, kg. Estimated bitumen stock. total_weight — Numeric, kg. Total estimated material stock, calculated as: total_weight = Steel + Wood + Cement + Stone + Sand + Brick + Mortar_plaster + Glass + Bitumen DATA INTERPRETATION The material fields are modeled material-stock estimates, not direct physical measurements. Material stock is the estimated mass of material associated with a record. It should not be confused with material intensity, which is normalized by floor area and may be expressed in kg m−2. Preserve kilograms as the source unit. Divide by 1,000 to convert kilograms to metric tonnes and by 10^12 to convert kilograms to gigatonnes. Document all conversions, filters, joins, reprojections, exclusions, and aggregation procedures. Zero, null, and missing values should be treated as distinct unless the associated manuscript or methods explicitly define them as equivalent. The nine material fields and total_weight may be summed across non-overlapping building records to obtain aggregate estimates. OBJECTID and geometry coordinates are not additive. Consult the associated manuscript for model assumptions, coefficient sources, building-function classifications, floor-height assumptions, spatial processing, validation, uncertainty, and the intended analytical scale. APPROPRIATE USES AND LIMITATIONS Potential uses include: • reproducing or extending analyses reported in the associated manuscript;• comparing material-stock patterns among cities or regions at the scale supported by the data;• examining relationships between urban form and estimated material stocks;• developing transparent scenarios for material-efficient urbanization; and• combining the estimates with other geospatial data after harmonizing spatial units, dates, definitions, and coordinate reference systems. Important limitations include: • estimates inherit uncertainty and bias from the source building data, building-function classification, material-intensity coefficients, assumed floor heights, geospatial processing, statistical methods, and model assumptions;• building-function misclassification may affect the assigned material coefficients and estimated stock of an individual record;• some records may represent merged adjacent buildings owing to preprocessing;• data completeness and accuracy may vary geographically;• modeled values should not be treated as cadastral, engineering, structural-safety, or direct building-survey measurements;• the dataset is not suitable by itself for site-specific engineering, regulatory, investment, or demolition decisions; and• comparisons may be misleading if spatial units, reference dates, building definitions, material categories, or denominators differ. LICENCE AND REUSE This dataset is released under the CC BY 4.0 licence. Users may share and adapt the data for any purpose, provided that appropriate credit is given, a link to the licence is supplied, and any changes are indicated. When redistributing transformed data: • retain the dataset DOI and creator attribution;• identify the source version;• document the transformations performed;• distinguish derived data from the original deposit; and• do not imply endorsement by the dataset creator. Users remain responsible for checking whether any third-party data combined with this dataset carry additional attribution or reuse requirements.