The first Chinese COVID-19 fake news dataset based on the Weibo platform.
We care about users' privacy and made (will keep making) efforts to protecting it.
- We did not make the
user name
public, which enables to identify Weibo users. In addition, we released the hasheduser_id
instead of the originaluser_id
of Weibo. - Please use the CHECKED data only for academic research.
This repository includes 2 folders which are code and dataset.
The code
folder contains codes for collecting COVID-19 microblogs and analyzing collected data.
The dataset
folder includes 2 folders - fake_news
and real_news
where
fake_news
includes 344 microblogs labeled 'fake'.real_news
involves 1776 microblogs labeled 'real'.
Each microblog (json file) consists of the following components:
id
: The microblog's ID. Each microblog is identified by a unique 16 digit ID assigned by Weibo.label
: The microblog's label. The label of each microblog is either 'real' or 'fake'.date
: The date that the microblog is posted (in format yyyy-mm-dd hh:mm).user_id
: The user's ID who posted the microblog. Weibo assigns each of its users a unique ID number with 10 digits. Note that each user can change his ID only once to a string formed by 4-20 characters, where letters are allowed.text
: Textual information of microblogs.pic url
: The URL of the visual information of microblog. Note that users are allowed to attach no more than 18 images in each microblog.video url
: The URL of the video information of microblog. Note that each microblog (i) can only include at most one video; and (ii) cannot attach both the video and image.comment num
,repost num
, andlike num
: The number of comments, forwards, and likes of the microblog.comments
: The detailed information of user comments for the microblog, including (i) the ID, date, and content of comments (microblogs), and (ii) the ID and name of commenters (users). Note that for each comment, no more than one image and no video are allowed.reposts
: The detailed information of user forwards for the microblog, which specifies (i) the ID, date, and content of forwards (microblogs), and (ii) the ID and name of forwarders (users). Similar to comments, each forward has at most one image and no video information. Note that if a user forwards a repost with an image, thepic_url
of the new forward will also include this image along with the original image.
news_fake.csv
contains all the fake microblogs in fake_news
with identical components except for 'comments' and 'reposts'.
news_real.csv
contains all the real microblogs in real_news
with identical components except for 'comments' and 'reposts'.
news_total.csv
is a combination of news_fake.csv
and news_real.csv
.
list.txt
includes all the keywords that we used to determine if the microblog about COVID-19.
If you are using this dataset, please kindly cite the following paper:
@article{yang2020checked,
title={CHECKED: Chinese COVID-19 Fake News Dataset},
author={Yang, Chen and Zhou, Xinyi and Zafarani, Reza},
journal={arXiv preprint arXiv:2010.09029},
year={2020}
}
Please contact [email protected] if you have any question on our dataset.