There are countless travel and booking sites which encourage travelers to write hotel, restaurant or attraction reviews. Some of these sites are international, while others cater to a local audience. Every day hundreds of thousands of reviews get published.
Olery scans millions of pages daily for new reviews, collects and stores them through various technologies in near real-time. We call this system the Olery Data Platform. Do you nee applications, reports or our API or SDK make use of it.
Most reviews consist of two parts. One is the guest’s written comment, the other one are ratings. Many sites ask the reviewer to evaluations topics like cleanliness or service by stars, points or percentages. There is no universally accepted set of ratings. Every site employs their own system of ratings.
Olery captures all ratings and combines them into 9 rating categories and more than 40 subratings. The volume of ratings and the diversity of the reviewers make these rating categories a very reliable data source which can be monitored day-to-day over long periods of time.
A lot of the experiences guests have in hotels are documented in the written review comments. Natural Language Processing makes it possible to extract these insights from reviews in 6 languages.
Natural Language Processing
Natural Language Processing (NLP) describes the computer activity of understanding human (written) language. The large number of unique languages and their complexity make this an extremely challenging task.
Olery has been part of a project funded by the European Commission to build a set of ready-to-use tools which can analyze text for different domains. The tools are open-source, however Olery was granted exclusivity for the hospitality domain.
It is our goal to steadily improve the reliability and precision of this technology for the hospitality industry. This way we can extract information from millions of hotel reviews and other sources.
OpeNER (Open Polarity Enhanced Name Entity Recognition) is a project funded by the European Commission under the FP7 (7th Framework Program). OpeNER’s tools are able to identify entities, detect opinions and to perform sentiment analysis from texts in Dutch, English, German, French, Spanish and Italian.