Technology and Tourists
Travelling is great, whenever we are travelling for pleasure or business. We encounter new people, different ways of thinking and new challenges.
These days, when we have access to a lot of information it’s hard to decide what we want and what we need when we are travelling and that is why new technical solutions have to be designed in order to keep us informed about our destinations.
Travel information systems are a great help on the road. If we think about TripAdvisor or Booking.com and how these sites helped us, it is easier to realise how important technology can be in the field of tourism. But tourists are often experienced people and they want everything to be personalised. Travel Recommender Systems were designed to tend to this exact need. These systems aim to create personalised recommendations for tourists regarding food, accommodations, transportation and entertainment.
This article presents an overview of a system designed to gather information about touristic attractions. The system is based on extracted information and makes recommendations based on user tastes and preferences.
An important challenge for current touristic recommender systems is delivering user information about potential points of interest which is both relevant and trustworthy. Tourists get more and more interested in finding information which matches their particular tastes, rather than general-level offers which assume the same set of preferences for everyone. This is in line with the long tail trend identified in consumer behaviour in various fields. To illustrate, let’s assume ten people intend to go to the seaside. The travel agent will ask them about their general preferences such as budget, available amount of time etc. and send seven of them to the standard, predefined promotional offer in Spain while the other three, for whom money is not a problem, will be sent to some more expensive hotels in Sardinia. A user-dependent offer is very unlikely to be made though.
When designing a personalised tourism recommender system, one major issue is gathering sufficient and well-organised information which allows for further filtering based on user-own taste and interests. At the same time, a high-performance information gathering system should allow an easy integration of new information sources and features. A promising solution to this problem is to employ ontologies and reasoners that allow the system to infer new information rather than retrieve what is already stored in a database. Ontologies organise information into a flexible hierarchy, easy to extend and to query, while reasoners infer new facts from the already known ones.
The goal of the system is to find the best-matching touristic attractions for a user, starting from a predefined profile which describes their preferences. To this end, the system relies on extracting information from general ontologies, as well as on parsing documents in order to extract information about the point of interest retrieved from a map system. The system was designed with a pragmatic view in mind, i.e., to select those technologies which seem the most promising for accomplishing the design goals. Therefore, the main contribution consists of the proposed design which is oriented towards available technologies and solutions, together with an architecture used to integrate them. The incremental data gathering process and architectural style allow the seamless adaptation of the design to several other technologies and information providers.
The system has a modular design that uses a map system, called Openstreetmap to extract touristic attractions around a geographical point and uses a general-purpose ontology called YAGO2 that stores information from Wikipedia and uses some parsing techniques to extract information from text written in English. It is able to identify persons, locations, organisations and other types of entities that can be used to describe a destination and helps a user make a decision if that place is good for them.
The greatest thing about this system is that it can gather information about places without specialised databases filled in by specialised staff but it knows how to search for things related to a geographical place.