Skip to content
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
alpha_shape
results
.gitignore
README.md
alldbpediapoints.py
batch_country_dataset_generator.py
cPlace.py
cReport.py
cSpinner.py
countrylist.py
doxygen_conf
geom_functions.py
vagueplaces.py

README.md

vague-places

Vagueplaces Generator using DBpedia SPARQL endpoint.

Jordi Castells 10 August 2012

This software is implemented as a final project of a Geoinformatics Master course at ITC Faculty of Geo-Information Science and Earth Observation. With Dr. ir. R.A. (Rolf) de By supervisor.

The main course file is vagueplaces.py

Abstract

Online gazetteers provide services that help to geocode locations. Given the name, the gazetteerwill attempt to provide coordinates for the place, usually in point information. Point informationis not always the best solution: urban areas, administrative units, or rivers are better describedusing polygons or lines. There is a clear need for region geocoding.In official administrative units, the situation is not critical. This kind of data is most of thetime available from online spatial data repositories. On the other hand, non-official boundariesare difficult to find and, old boundaries or folklore delimitations difficult to obtain. For example,Mittelland in Switzerland or the Scottish highlands.Solutions regarding the web are already proposed by Arampatzis et al. and Christopher B. Jones et al. inside the research by the SPIRIT project. Thoseapproaches use google search and classification to obtain the information; this project will aim toobtain similar results using semantic web approaches.

Previous work presented a solution to the problem. The main aim of this small project is to provideanother way to obtain a similar or better solution with the aid of the semantic web formats using DBpediaas a base data repository

Requirements

To run vagueplaces.py you will have to provide a couple of libraries: shapely SPARQLWrapper

Running

Run vagueplaces.py --help to get a list of options. But mainly you choose which DBPedia to point (live or last release), a set of words to filter from and an alpha shape value.

Extra

There's a website presenting the points from DBPedia. To generate all the point dataset there are a couple of scripts that automatically handle the task.

alldbpediapoints.py will download all DBPedia entries with latitude and longitude.

countrylist.py takes an input file (generated from alldbpediapoints.py) and prints a list of "valid" country resources. Since the CSV may have strange artifact entries (wront lines, dirty lines) this script only takes into account valid lines and countries with more than 3 points.

batch_country_dataset_generator.py will take an input csv file generated by alldbpediapoints.py and split it in smaller files, one for each valid country.

Official Report

The official report can be found in PDF format either in Scribd or Mendeley:

http://www.scribd.com/doc/111990720/Defining-vague-places-with-web-knowledge http://www.mendeley.com/profiles/jordi-castells/

You can’t perform that action at this time.