Product Review Datasets: Epinions and Ciao


In recent years, the notion of trust has attracted more and more attention from community science community. Triditional research about trust is usually conducted in the physical world while it is very challenging to study trust in the online world since most of information that a sociologist collects for trust may not be available.

An important characteristic of product review sites such as Epinions is that there exist trust networks among users. Such sites provide a sensible platform to study trust in an online world. we crawled two datasets from two popular product review sites Epinions and Ciao in the month of May, 2011. More details about these datasets are described below respectively.

Epinions Dataset

For each user, we have his profile, his ratings and his trust relations. For each rating, we have the product name and its category, the rating score, the time point when the rating is created, and the helpfulness of this rating. Here we only release parts of the dataset, used in [1] with matlab forms. Its remaining parts will be released later or are available upon your request. If you want the raw dataset(.txt without preprocess), please feel free to contact with me.

Ciao Dataset

This dataset contains all the infromation Epinions has except the time points when the trust relations are established. However, we crawled the text content of each review and the helpfulness votes for this review from other users. Similarly, we only release parts of the dataset, used in [1] with matlab forms. However, other parts are available later or upon your request.



Updated on 10/23/2012