Xponents

Geographic Place, Date/time, and Pattern entity extraction toolkit along with text extraction from unstructured data and GIS outputters.

View the Project on GitHub

Gazetteer Curation and Data Model

SEE GAZETTEER_REPORT for source information and data rights.

Xponents SQLite ETL Schema

The Xponents SQLite schema represents the core ETL data to produce the converged and merged Gazetteer. From here it is indexed into Solr as the production gazetteer/tagger.


sqlite TABLE placenames (

	id              -- integer row ID used here and in Solr index 
	place_id        -- source place ID        
	name	        -- the place name, abbreviation or code. Any text, really	
	name_type       -- type of name N, A, C, respectively        
	name_group	    -- coarse grouping by language script/tokenization model. 
	                  general (""), Arabic ("ar"), and Chinese/Japanese/Korean ("cjk")	
	source	        -- source catalog ID
	feat_class      -- geonames-based feature class 
	feat_code       -- geonames-based feature coding
 	cc              -- ISO 3166 country code (2-alpha) or "*"
	FIPS_cc         -- FIPS 10-4 country code (2-alpha) or "*"
	adm1            -- ISO-based Admin Level 1 boundary code. FIPS enumerations are remapped to US postal ADM1
	adm2            -- Admin Level 2 boundary code. Not always provided
	lat, lon        -- WGS84 decimal latitude/longitude 
	geohash         -- 6-character geohash
	duplicate       -- distinct base names are default as duplicate=0; 
	                   duplicates on the KEY=<country+name+feature+geohash> are marked as duplicate=1
	name_bias       -- Internal name bias from -100 to 100 indicating if a name is worth tagging or for searching only
	id_bias         -- Internal place ID bias indicating *our* preference for this particular name+location pair 
	                   when that name appears 
	search_only     -- Flag to indicate this geography should be carried along with Gazetteer, but not used for tagging


INDICES on columns above help speed curation and Solr indexing:

*  place_id
*  source
*  cc
*  adm1
*  name
*  name_type
*  name_group
*  feat_class
*  feat_code
*  duplicate
*  search_only


Geonames

Original Geonames Schema is available at http://download.geonames.org/export/dump/readme.txt

NaturalEarth

A sampling of Natural Earth 10m boundary shapefile is printed here. Still evaluating which resolution file to use.

# As printed out using _schema()
NATEARTH_SCHEMA = """
DeletionFlag C
featurecla C
scalerank N
adm1_code C
diss_me N
iso_3166_2 C
wikipedia C
iso_a2 C
adm0_sr N
name C
name_alt C
name_local C
type C
type_en C
code_local C
code_hasc C
note C
hasc_maybe C
region C
region_cod C
provnum_ne N
gadm_level N
check_me N
datarank N
abbrev C
postal C
area_sqkm N
sameascity N
labelrank N
name_len N
mapcolor9 N
mapcolor13 N
fips C
fips_alt C
woe_id N
woe_label C
woe_name C
latitude N
longitude N
sov_a3 C
adm0_a3 C
adm0_label N
admin C
geonunit C
gu_a3 C
gn_id N
gn_name C
gns_id N
gns_name C
gn_level N
gn_region C
gn_a1_code C
region_sub C
sub_code C
gns_level N
gns_lang C
gns_adm1 C
gns_region C
min_label N
max_label N
min_zoom N
wikidataid C
name_ar C
name_bn C
name_de C
name_en C
name_es C
name_fr C
name_el C
name_hi C
name_hu C
name_id C
name_it C
name_ja C
name_ko C
name_nl C
name_pl C
name_pt C
name_ru C
name_sv C
name_tr C
name_vi C
name_zh C
ne_id N
"""