Home

Awesome

B2W-Reviews01

B2W-Reviews01 is an open corpus of product reviews. It contains more than 130k e-commerce customer reviews, collected from the Americanas.com website between January and May, 2018. B2W-Reviews01 offers rich information about the reviewer profile, such as gender, age, and geographical location. The corpus also has two different review rates:

Applications

This corpus can be useful for several Natural Language Processing (NLP)/ Computational Linguistics (CL) tasks. The first that comes to mind is probably sentiment analysis. Sentiment analysis is the task of assigning a sentiment (or a position) to the content of a given text. For this task, B2W-Reviews01 offers the two distinct evaluation ratings.

Product reviews often have complex information, related not only to the product that was purchased, but also to the online shopping experience, payment methods, or even the product delivery process. Therefore, different facts and opinions can be extracted from such a corpus, and classifying sentiment may not be enough to capture the content of reviews. For real world applications, dealing with topic modeling, user intent identification and feature extraction also become necessary. It is relevant to know not only the reviewer’s sentiment, but also the object of this feeling.

Since B2W-Reviews01 offers the exact text written by users, this corpus also offers rich material for those interested on out-of-vocabulary words, slang identification, or spell-checker tasks. For those interested on socio-linguistics analysis, the present corpus offers a rich possibility of crossing reviewer information considering gender, age and geographical location. One can, for example, find easily how negative or positive reviews are distributed among age groups or which product categories receive more reviews from women or men. It is also possible to conduct a study on bias in reviews by joining and aggregating data.

Although B2W-Reviews01 is mainly a product review dataset, we believe that important insights about the current language in use in the web register can be made, since Americanas.com customers are spread throughout Brazil and have different social backgrounds.

CSV file in a spreadsheet software

To correctly import the B2W-Reviews01.csv file into a spreadsheet software, make sure to mark only semicolon (;) as the delimiter and use double quotes (") as text qualifiers.

Licensing

B2W-Reviews01 is available at https://github.com/b2wdigital/b2w-reviews01 under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.01, https://creativecommons.org/licenses/by-nc-sa/4.0/), which means that licensees may only copy, distribute, display, work on and make derivative works and remixes based on it if they give credit to B2W Digital in the manner specified in this work. Also, licensees may only distribute derivative works under a license identical (“not more restrictive”) to the license that governs the original work. Finally, licensees may only copy, distribute, display, work on, and make derivative works and remixes based on it for non-commercial purposes.