World-Islands: the data behind the map

It started with a simple question on a crowded map: how many islands are there, really, and what stories do they tell when you line up climate, coasts, ports, airports, mountains, seas, sunlight, and people? We took the long way around. I extracted global geo-data, cleaned it until it behaved, and shaped it into something a human (and a researcher) can work with—searchable, joinable, and ready for analysis or storytelling. I’m happy to share curated exports for scientific use.

What’s inside the dataset

The dataset covers islands worldwide, from large landmasses to smaller islets where sources allow. Each island is represented by a polygon and a centroid in WGS84, with generalized zoom levels for the web. Core attributes include names and variants; country and admin affiliation (where applicable); area and coastline length; Köppen–Geiger climate class; nearest port (with World Port Index ID) and distance; airport presence and nearest facility from OurAirports; mountain context (highest point where available, plus a “mountainous” flag drawing on hierarchical inventories and peaklist.org); IHO sea area tags from VLIZ; population density overlays from CIESIN GPWv4; and solar resource summaries from the Global Solar Atlas. Typical uses range from comparative island studies and logistics to climate, energy, and education.

How I built it

I compiled everything in an SQL database, using the QGIS toolset for ingest, processing, and QA.

Ingest
Upstream vectors and rasters were loaded in their native formats with version and date captured in metadata.

Harmonize
All geometries were stored in WGS84 for interoperability; area and coastline metrics were computed in an equal-area projection to avoid distortion.

Clean
We ran topology checks, dissolved slivers, handled the antimeridian, and standardized names and Unicode. Where sources disagreed on naming or grouping, we retained multiple variants.

Join
Spatial joins and nearest-neighbor lookups appended climate class, ports, airports, mountains, sea areas, population, and solar metrics to each island record.

Aggregate
We derived fields such as area_km2, coastline_km, climate_code, nearest_port_km, pop_density_mean, solar_gti_mean, and a mountain_flag; where available, we added highest_peak_m and peak_name.

Export
For researchers, we offer GeoPackage/GeoJSON (full geometry) and CSV/Parquet (attributes + centroid), plus lightweight web tiles for previews. Each release ships with a README and data dictionary.

Publish

The result is published on world-islands.net

Methodological notes and caveats

Scale and resolution matter. Very small islets are sensitive to source scale; coastline estimates on tiny polygons carry higher uncertainty. Köppen–Geiger at ~1 km and population rasters can over- or under-estimate on small footprints. Each upstream source follows its own update cycle; attributes reflect the latest ingested versions listed in the metadata. Archipelago membership and local names vary by source; we keep alternatives rather than force a single truth. Licenses remain with their original owners; derived aggregates are shared subject to those terms—please review original licenses before redistribution.

Access for scientific use

If you’re a researcher or educator and would like the data, tell me your affiliation and intended use. Typical shares include the full attribute table, centroids, and—on request—polygons for specific study regions. Available formats are GeoPackage, GeoJSON, and CSV/Parquet. Citation guidance appears below; please acknowledge upstream sources in any publication.

Suggested citation

“World-Islands (compiled dataset). Version [YYYY-MM]. Aggregated from publicly accessible sources including Global Islands (Sayre 2022), Natural Earth, Beck et al. (2018), OurAirports, World Port Index, CIESIN GPWv4, GEBCO, VLIZ IHO Sea Areas, Global Solar Atlas, Snethlage et al. (2022), and peaklist.org; processed in SQL with QGIS.”

Acknowledgments and sources

Special thanks to the organizations and individuals who have contributed to the collection and publication of these publicly accessible data. A complete list of sources and acknowledgments can be found here.

Datasets & publications

R. Sayre, “Global Islands.” GEO Knowledge Hub, June 10, 2022. doi: 10.60566/7yjze-2g558
Made with Natural Earth. Free vector and raster map data @ naturalearthdata.com
Beck, H., Zimmermann, N., McVicar, T. et al. Present and future Köppen–Geiger climate classification maps at 1-km resolution. Scientific Data 5, 180214 (2018).
OurAirports dataset.
National Geospatial-Intelligence Agency. 2019. “World Port Index.”
Snethlage, M.A., Geschke, J., Spehn, E.M., Ranipeta, A., Yoccoz, N.G., Körner, Ch., Jetz, W., Fischer, M. & Urbach, D. A hierarchical inventory of the world’s mountains for global comparative mountain science. Nature Scientific Data (2022). https://doi.org/10.1038/s41597-022-01256-y
Mountains over 1500 m by peaklist.org
VLIZ (2005). IHO Sea Areas. Available at marineregions.org
Global Solar Atlas 2.0, developed and operated by Solargis s.r.o. for the World Bank Group with funding from ESMAP. https://globalsolaratlas.info
Satellite imagery basemap: NASA Earth Observatory map by Joshua Stevens using data from NASA’s MODIS Land Cover, the Shuttle Radar Topography Mission (SRTM), the General Bathymetric Chart of the Oceans (GEBCO), and Natural Earth boundaries.
GEBCO — General Bathymetric Chart of the Oceans.
Center for International Earth Science Information Network (CIESIN) – Columbia University. 2018. Gridded Population of the World, Version 4 (GPWv4): Population Density, Revision 11. NASA SEDAC.

Ethics and responsible use

Respect indigenous place names and local conventions where available. Do not use population overlays to target vulnerable communities. When in doubt, prefer transparency: methods, thresholds, and exclusions should be documented.

FAQ

What is the smallest island included?
Inclusion depends on source resolution; very small islets may be generalized or absent.

Can I get only attributes without polygons?
Yes. Centroid-only exports and attribute tables are available for lightweight work.

Do you provide an API?
Data are shared as files for now. If you need programmatic access for a study, get in touch.