The files in this directory represent the results of a spatiotemporal mapping of Wikipedia concepts. They were extracted using Wikipedia dumps and a DBPedia geotagged articles list available in July 2009. Spatial mapping is performed by analyzing links to geotagged Wikipedia in up to 7 article versions (English, German, French, Dutch, Spanish, Italian and Portuguese) and by ranking first the locations which are linked in the highest number of different languages. Temporal mapping is limited to Wikipedia biographies and it involves and extraction of a person's approximate life period (within the right century/centuries). All extractions are automatic and there is no guarantee concerning their correctness.

A detailed description of the mapping process is available in “Adrian Popescu, Gregory Grefenstette Spatiotemporal Mapping of Wikipedia Concepts, JCDL 2010, June 21 - 25, Brisbane, Australia”.

The following files are available:


-dbpedia_locations.txt.zip - contains an enriched list of geotagged DBPedia articles. The file has the following structure:
	CONCEPT\tLATITUDE\tLONGITUDE\tTRANSLATIONS\tCATEGORIES\tCOUNTRY\tCITY

where:
 	-\t - tabulation
	-CONCEPT - title of the English Wikipedia article that describes a concept
	-LATITUDE - latitude of the location
	-LONGITUDE - longitude of the location
	-TRANSLATIONS - links to  German, French, Dutch, Spanish, Italian and Portuguese versions (if these versions exist and are linked in the English version). Separation is done using '; '.
	-CATEGORIES - categories of the English Wikipedia article. Separation is done using '; '.
	-COUNTRY - country where the location is situated (if applicable)
	-CITY - city where the location is situated (if applicable). Cities are considered self-included.
	NB: some location names (CONCEPT) and city names (CITY) are linked under multiple names (for instance Los Angeles and Los Angeles, California). In such cases alternative names both appear in the CONCEPT and CITY field separated by three white spaces

-all_wikipedia.txt.zip - spatial mapping of all articles in English Wikipedia. The file has the following structure:

CONCEPT\tCATEGORIES\tTRANSLATIONS\tLOCATIONS

where:
	-\t - tabulation
	-CONCEPT - see above
	-CATEGORIES - see above
	-TRANSLATIONS - see above
	-LOCATIONS - list of locations associated to CONCEPT. If several locations are discovered, they are separated by '   ' (three white spaces). Each location has the following structure

	NAME; LATITUDE; LONGITUDE; R1; R2; R3; COUNTRY; CITY

	where:
		-NAME - location name
		-LATITUDE - see above
		-LONGITUDE - see above
		-R1 - the number of different languages NAME is linked in. The maximum value is 7 (the number of languages that were processed)
		-R2 - the total number of links to NAME in all article versions
		-R3 - the total number of mentions of NAME in all article versions.
		-COUNTRY - see above
		-CITY - see above
	NB: some location names (NAME) and city names (CITY) are linked under multiple names (for instance Los Angeles and Los Angeles, California). In such cases alternative names both appear in the NAME and CITY field separated by two white spaces
 

-people_space_time.txt.zip - is a subset of all_wikipedia.txt which contains only Wikipedia biographies for which it was possible to extract temporal information. The structure of the file is the following: 
	
CONCEPT\tCATEGORIES\tTRANSLATIONS\tLOCATIONS\tTIME
where:	
	-time - the century/centuries when the person lived