Is there a source for historical shapefiles and tiles? Like the ancient Greece/Rome, medieval Europe and Japan? Ideally with multiple snapshots during different periods in their development, since they've gone through quite a bit of restructuring over their lifespans.
take a look at: http://nils.weidmann.ws/projects/cshapes
Quote from the website: "CShapes is a new dataset that provides historical maps of state boundaries and capitals in the post-World War II period. The dataset is coded according to both the Correlates of War and the Gleditsch and Ward (1999) state lists, and is therefore compatible with a great number of existing databases in the discipline. Provided in a geographic data format, CShapes can be used directly with standard GIS software, allowing a wide range of spatial computations. In addition, we supply a CShapes package for the R statistical toolkit. This package enables researchers without GIS skills to perform various useful operations on the GIS maps"
Update: Version2 of Cshapes can be found here: https://icr.ethz.ch/data/cshapes/
Here’s something for the US.
The National Historical Geographic Information System (NHGIS) provides, free of charge, aggregate census data and GIS-compatible boundary files for the United States between 1790 and 2010.
Great Britain has http://visionofbritain.org.uk/
Using Scanned Maps that have been georeferenced and combined with historical elements, census, election results and more.
Example - Ordnance Survey Unions 1803
more info on the project: http://visionofbritain.org.uk/index.jsp
I don't have a link for data but this use case is extremely interesting.
coincidental post on another site.
Q: Just for the sake of curiosity Tom, what type of projects are you doing with such a powerful machine?
A: "Glad you asked. The primary use is in attempting to locate the precise camera location of a set of exposures taken somewhere on the Gettysburg battlefield one day after the battle. The ArcObject program I wrote attempts to do this by matching the 2 dimensional horizon visible in one of the exposures with that visible from every 10' x 10' location on the historical battlefield. Rays are projected toward the horizon at every 4 degree increment. Using a sort algorithm the inclination and distance to the horizon is determined. From this a 2-D horizon is calculated and match attempted... The result is 90 10'x10' rasters with a value that reflects the likelyhood of matching. A total run takes just 30 days using all the processors. There is also a spatial model that uses 1863 Shapefile to reflect other elements visible in the exposures (water, slope, ground cover, etc.). This model produces 90 additional raster that are combined with the first set... Ultimately producing suggested camera locations and viewing directions; which require field trips to evaluate and update the models... "
Re-post from a linkedin group. Now that is digging into history!