Travel Semantics
Snippets, Semantics and the Web 3.0: structured and meaningful Data in
Online Travel Marketing and Technology
Dr. Lars Göhler
o Wilhelm von Humboldt: „Language makes
infinite use of finite means.”
o Tim Berners Lee: "The goal of the
semantic web is to express real life.”
Semantic Technologies:
The Situation
o Big and inceasing unstructured web data in travel and tourism offering and advertising make a sensible mining, search and selection almost impossible.
o Structured data could make classification and ordering much easier.
o With the application of semantic methods, those data can even be „understood“ by machines and complex results yet unexpected can be produced.
Tim Berners Lee:
Creator of the World Wide Web
„Hi, I invented the world
wide web!“
„It can be much more
powerful as the
semantic web.“
„Link your data, give
them a structure and
a meaning!“
Photographer: John S. and
James L. Knight Foundation
For us: a beautiful website
For a machine like Google:
a chaos of information
Photo: Claudia Hautumm / pixelio.de
What Machines do now:
Comparing Structures
structure 1 structure 2
apache apache ? Webserver? American Indian? Helicopter? Car?
Example:
What machines can do:
comparing structures with reference to an ontology
Structure 1 Structure 2
concept
sign and
syntax
level
meaning,
ontology
level
Semantic Approach:
Example
apache apache
Concept
of a web-
server
sign and
syntax
level
meaning,
ontology
level
reality
Semantics can do more:
Understanding Relations
Yucatan Paul likes to travel to
…and also Complex Relations,
even between Ontologies
source: linkeddata.org
Types of references:
low semantics
o Confined to logical HTML-tags as <nav>,
<article> or <footer>
o Mainly self-referencial
o Marking the structure of a website, hardly
any external reference
o Encouraged in HTML5, but this standard is
more about the capacity to include
semantic markup into HTML-code
Types of References 1:
Self Reference (HTML 5)
Hi, I‘m a navigation!
Image source: expedia.de
Types of references 2:
External linked Reference
Hi, I‘m an aggregate offer
according to schema.org!
Image source: weg.de
Types of references 2:
medium semantics
o Reference to an external established source or
ontology as schema.org or data-vocabulary.org
o Uses simple rdf or microdata
o advantages: public, accepted by most search
engines, well known, easy to interpret, honoured
by search engines by „rich snippets“
o disatvantages: no subject specific, limited to the
available vocabulary, limited capacity to express
relations
Types of references 3:
full semantic reference
Hi, I could be a complex hospitality offer
according to an travel ontology xxx
(that does not yet exist).
Image source: weg.de
Types of References 3:
Full Semantics
o external linked reference to a complex
ontology
o Uses RDF(a) instead of microdata or
microformats
o able to express various relationships in the
form of triples
o difficulty: presupposes ontologies that are
widely accepted
Use of semantic Technologies:
Search
o If linked to an ontology we can give
expressions a unequivocal meaning
o We can link synonyms to the same
meaning and homonyms can be attributed
a special meaning.
o We can express relationships between
those meanings
Use of semantic Technologies:
Internal Search
o Users can search for journey in everyday language:
o We can link synonyms to the same meaning and homonyms can be attributed a special meaning.
o We can express relationships between those meanings
o Search algorithms can react on complex requests: „I want a best rated hotel in the center of Olso with a strong wifi connection and a conference room.“
Use of semantic Technologies:
Internal Search
yourvisit.com (search technology by Yahoo)
zaptravel.com
Use of semantic technologies:
Search with Fact-Finder
neckermann-urlaubswelt.de (using search technology from fact-finder.de)
answer:
Perspectives:
Internal semantic Search
o Extending the ability of Websites and mobile devices to understand complex queries in natural language
o Acoustic speech recognition a semantic basis, especially for mobile devices
o Making travel systems understand complex requirements, e.g. „family trip“, „golf journey“, not only time, place and price.
o It is said, that semantic searches have a fare greater conversion rate.
Status:
Semantic Markup in Travel Websites
o Few metadata are used by almost every
company, describing the company and the
breadcrumb
o Germany: weg.de and some sites of unister
travel use of microdata for single offers
o More complex markup, e.g. references to Good
Relations is hardly to find.
o Résumé: Almost no systematic metadata
markup is applied in most travel websites.
Status:
External Search
o Almost all search engines, especially Google „honour“
semantically marked offers by a special presentation,
called „snippet“ (search for Reise Antalya).
Use of semantic search technologies 2:
Knowledge Graph Knowledge graph is a knowledge base built by Google to enhance search
results by semantic search information using several sources
Perspectives:
Public Search, SEO
o enhance search engine presence by marking up
offers with metadata
o enable search engines to classify data as place,
offer, event, transportation, accommodation etc.
o describe media as videos and photos by
metadata to enable search enginges to
„understand“ their content
o contribute to information systems as knowledge
graph etc.
Perspectives:
Social Media
o Semantic technologies can interact mit
social media
o It is technically possible to search for the
„hotel in Paris with the best rating“
o Most social media organize community
data with semantic technologies, e.g.
Facebooks open or social graph and even
Google+
Status:
Destination Marketing
o Most destinations, places or offices are
registered in google maps/places
o destinations start to discover the
opportunities of publishing geodata
o Hardly any geodata are linked to other
ressources (booking, events, associated
offers, packages)
Perspectives:
Destination Marketing
o Publication of semantic data for destination
marketing in an universally accepted form
o Publishing Metadata for tourist services and
facilities
o Link these data to each other and to other data
of transport, flights, events etc.
o Mapping those data in applications using Open
Street Map, Google Maps, Earth etc.
o Creating special maps for tourist destinations
Status:
Mobile Solutions
o Tripit (tripit.com): online/mobile travel itinery
planner that uses semantic technologies
o Trust Score (trustyou.com) big data online
reputation management: collects hotel
reviews from many web sources on a
semantic basis
o Desti is a iPad app: semantic search travel
companion (yet only available for iOS)
Perspectives:
Mobile Solutions
o Use of location based services in
tourism and travel is still in its infancy
o Native and WebApps can use geodata
and mix them with social media, events,
user rating and semantic search
o Regions and destinations can create
own travel companions with many
locally relevant data
Status and perspectives:
Semantic Webservices
o Many tourist and travel companies (Sabre, Amadeus etc.) offer numerous Webservices to which any agent with the proper interface can connect
o Most of these Webservices are non-semantic, some of them use the OTA Vocabulary
o Perspectives: Semantic Webservices could offer a far more flexible search and offering system, even connecting more than one service. No special interface specifications needed, any semantic agent can connect
Status:
Infrastructural Solutions
o The Open Travel Alliance created a complex
structured system of travel vocabulary, but only
partly transferred to an ontology.
o schema.org: small ontology, with hierarchy and
nesting possibilities, no triples, no extensive use
of relations
o Good Relations: web vocabulary for
e-commerce; the only complex public ontology
supported by Google, and Yahoo
Available Solutions:
Open Travel Vocabulary (XML-based)
Perspectives:
Internal Solutions
o Creating company-intern ontology for travel and tourism would have positive effects to all search and data mining functions
o Can be extended to an internal semantic data organization as a means to coordinate all internal data formats
o Internal ontologies can be mapped/related to accepted external ontologies
Perspectives:
Infrastructural Solutions o Creating an widely accepted (leightweight) public
ontology for travel and tourism with a good hierarchy would have advantages for the user, for companies and for third party applications
o Every application can use the published public data without a special interface: the data could get a wider recognition
o Can yield much more appropriate search results in search engines
o If computers can „understand“ search queries and available data alike, they can construe complex answers and offers.
Perspectives:
Strategic Questions
o Should companies develop own ontologies or concentrate on using established ones?
o Is it better to follow the major search engines or develop an own strategy, say, oriented on open standards?
o Will semantic technology be an additional system to manage heterogenous data or does it make sense to develop it as a core technology?
Presentation held at the ITB Berlin 2014
by Lars Göhler
on March 6th, at 11.30 – 12:00
at the Hall 6.1, eTravel Lab
www.eastpress.de
Further information on travel semantics on:
www.travel-semantics.com