Page 397 - NGTU_paper_withoutVideo

Modern Geomatics Technologies and Applications

Semantic Approach to Spatial Information Retrieval - Case Study: Vancouver
Downtown Traffic Nodes During Rainfall

1
Mohammad Hadi Yazdian and Mahmoud Reza Delavar 2*

1 GIS Department, School of Surveying and Geospatial Eng., College of Engineering, University of Tehran,
North Kargar Ave., Jalal Al Ahmad Crossing Tehran, Iran,
hadiyazdian@ut.ac.ir
2 Center of Excellence in Geomatic Eng. in Disaster Management, School of Surveying and Geospatial Eng.,
College of Engineering, University of Tehran, North Kargar Ave., Jalal Al Ahmad Crossing Tehran, Iran,
* mdelavar@ut.ac.ir

Abstract: In modern cities with high traffic congestion, rainfall has always been one of the main causes of disruption of
the urban traffic system. On the other hand, data processing and analysis, especially for spatial data, today is more limited
to separate and independent data sets that need to make correlations among related entities to solve complex problems
and perform various processes. Previous studies such as data mining on linked data, search and retrieval spatial data and
traffic density monitoring during rainfall have been performed، but due to the many hours without rainfall, excessive data
is stored that caused data redundancy and the accumulation of worthless data, but the use of spatial and non-spatial data
available in different sources with values and information Differentiation and conversion of data into triple graph model
in the use of semantic web is one of the activities that has been less considered in previous researches. In this study, while
investigating the effect of rainfall on traffic flow using linked data to Let's take a look at the advantages of choosing this
technology. We use Vancouver's rainfall and traffic data for two weeks . with using semantic web and related technologies
by converting web-extracted data into RDF data model And link them to a database We retrieved data using
GEOSPARQL technology. In this study, using the RDF triple data model and storing data in a base graph database and
using semantic-spatial concepts, we observed that this method has the ability to provide solutions for using different
sources of information. In addition, we were able to prevent the entry of valueless data, which prevented the creation of
dependencies in the model and the entry of worthless data, which maintains the consistency of the model and prevents any
complexity from entering the model.

Keywords: Linked-data, sematic-web, GeoSPARQL, Traffic, Query language, geospatial data, RDF

1. Introduction
1
With the introduction of the concept of the Semantic Web in 2003 by the W3C, a new approach to the design of semantic
standards and models began with the aim of converting the content of web pages to use web-based data. Semantic Web means
that when computers, in addition to retrieving data, have the ability to understand the data available on the Web, enabling the
users to have new types of Web and intelligent Web applications[1]. The first and most tangible result of creating a semantic
web will be the evolution of information retrieval. Web data retrieval is generally based on matching the words and phrases
searched with the words and phrases in the text of the web pages. The Semantic Web goes beyond matching words and searches
by subject, data link, data type, and other qualities.

Until recently, Web was only considered as an information space for interrelated documents in which some documents
are linked. The type and nature of the connection between the documents are unclear, and in this case a major part of the structure
and meanings was sacrificed. In recent years, the Web has evolved from a global space where documents are interlinked to a
space where data in addition to documents are interlinked. This evolution is a set of efforts to publish and connect structured
data on the web, known as linked data [2].
With the increasing number of spatial data, we are faced with a large volume of this type of data that the need to reconsider
how to store, use and decide on the choice of a particular data set, has become more urgent.
Spatial data is available in such a way that it is either directly observed by various measurement tools or generated as
output during the modeling and computational process. These different trends have led to the heterogeneity of existing spatial

1 World Wide Web Consortium

1