Archive for geodata

Useful geodata links

Posted in Data with tags , on February 22, 2010 by hansvandermaarel

I’ve been asked recently about sources of useful (free) geodata. Since I thought it might be nice for other people to have as well I’ve decided to make a blog post about it. So here goes.

Natural Earth Vector – Deserves to be mentioned first. A global cartographic (!) dataset at scales 1:10M, 1:50M and 1:110M, both vector and raster. Made possible by Tom Patterson, Nathaniel Kelso and a small army of volunteers. Made by cartographers, for cartographers.

Geocommons Finder – Offers a wide range of thematic datasets.

GIS Data Depot – Requires registration but again a good source of data. VMAP / DCW per country, many US datasets per county.

SRTM data – Elevation data in 90 meter resolution (30 meters in the continental US) data in 1×1 degree tiles, or go to the CGIAR site for corrected data in 5×5 degree tiles. SRTM only covers the earth from 60 North to 60 South and does not offer bathymetry. Viewfinder offers manually corrected data for significant mountain areas as well as manually created data for areas not covered by SRTM.

Aster Global DEM – Higher resolution than SRTM, but I’ve heard mixed reports about the data quality.

MapStor – Old USSR topographic maps, either free (download one by one) or dirt cheap even for large areas. While most of these maps are by now quite outdated (they were, after all, produced before the collapse of the Soviet Union), in many parts of the world they are the only easily obtainable source of reference data. If you speak Russian of course… I’ve personally used this to track down villages in India that Google doesn’t show. The maps come with a georeference in OziExplorer format. Global Mapper can read this and can also automatically remove the collars.

Global Land Cover Facility – Offers various raster datasets, including Landsat, for download. Much of the data is offered in seperate bands, so you’ll need to do a bit of research to use it.

Geonames – USGS board on geographic names. Note that the data is not without errors, so always doublecheck.

True Marble – 250 meter (free) and 15 meter (paid) satellite data. Automatically generated out of Landsat data. Some artifacts here and there but overall a good dataset.

David Rumsey collection – Not geodata per-se, but a vast collection of old maps.

Perry-CastaƱeda Library – Large collection of reference maps offered by the University of Texas.

Cloudmade and Geofabrik offer shapefile extracts from the OpenStreetmap database by continent/country/state.

EarthExplorer and the Seamless Server are two USGS sites offering data for download. Mainly US data as far as I can tell.

US TIGER – vector data of the US offered by the US Cencus Bureau. Its quality is not always 100%, see this post on Cartotalk for an explanation.

US National Atlas – Many thematic datasets for the US.

Natural Earth Vector

Posted in Data, FME with tags , , on July 16, 2009 by hansvandermaarel

Ever since Tom Patterson started offering Natural Earth, a public domain raster dataset of the entire Earth, I’ve been a big fan of it. In fact, the base layer for the Oolaalaa globes is based on Natural Earth I (with added bathymetry and increased contrast).

After releasing 3 raster datasets this way, Tom has recently teamed up with Nathaniel V. Kelso to create a vector product that also will be released into the public domain: Natural Earth Vector. Nathaniel has invited me to participate in this as well, and to be honest I think it’s a wonderful project. So far I’ve been involved in extracting road data from OpenStreetmap and identifying rivers in Europe.

Processing the OpenStreetmap data, using FME, reinforced my slight aversion of XML/GML based data. When downloading the data, you have the option of getting the .osm data, which is in an XML format, or shapefiles. FME does support the .osm format, but processing the files takes a terrible amount of time. Processing the shapefiles for Europe, Asia, Oceania, Africa and South America, just filtering out the major highways, takes about 11 minutes. I don’t dare to do that with the .osm data, but I fear it may take days. It’s nice that XML is so open and free, but for large amounts of data it’s simply unsuitable.

Anyway, to get back to Natural Earth Vector, it will be officially unveiled at the NACIS conference in Sacramento this october and I for one am looking very much forward to that.