Crawled over 2 weeks in January 2014, the Webis TripAdvisor Corpus 2014 (Webis-Tripad-14) consists of 266 061 reviews on 12 044 hotels by 208 785 users. Additionally, there is meta data about the hotels (such as location or overall ratings), the users (such as gender and age range) and the reviews itself (such as date posted and rating) available. We offer a download in json format: one file per hotel and one file containing all the user information.
To download the corpus use the following link:
(61 MB, MD5 sum: e103d1c66efbbed46159e4225d17ad9f)
The Webis TripAdvisor Corpus 2014 (Webis-Tripad-14) is designed in such a way that several different tasks can be performed on it, such as sentiment analysis, author profiling or usefulness detection.
The json-corpus consists of 12 045 files, where one of them contains all the user data and the others are one for each of the hotels in the data set. A detailed description of the data and the key/value pairs can be found as a README.txt in the download folder.
Useful statistics are shown in the tables below.
|Attribute||Number of Entries|
|No gender given||119393|
|13-17 years old||Total users||108|
|No gender given||2|
|18-24 years old||Total users||2220|
|No gender given||32|
|25-34 years old||Total users||22286|
|No gender given||436|
|35-49 years old||Total users||37610|
|No gender given||735|
|50-64 years old||Total users||24595|
|No gender given||417|
|No gender given||53|
|No age given||117718|
|Reviews per hotel||Min||1|
|Words per review||Min||0|
|Reviews with helpfulness votes||154291|
|Reviews without helpfulness votes||111770|
|Votes per review with votes||Min||1|
Students: Katharina Spiel