Weverse - BTS Feed Data

This dataset contains user information and text data (posts and comments) gathered from Weverse (https://weverse.io/bts/feed).

Using Python’s Selenium package (https://www.selenium.dev/), we crawled the Weverse BTS channel (https://weverse.io/bts/feed) on March 3, 2024, starting at 11:00 p.m. and continuing for two hours (from 11:13 p.m. on March 3 to 1:26 a.m. on March 4). This procedure yielded 16,020 posts and 14,223 unique user IDs.

Because our aim was to investigate the behavior of established, active users rather than recent joiners, we paused data collection for a few months. On May 24, 2024, we revisited the Weverse BTS channel and accessed each previously identified user ID via its profile page at https://weverse.com/bts/\{profile_id\}. We then crawled all posts and comments these users had written during the two-month period from March 3 to May 3. Profiles set to private or belonging to deleted accounts were inaccessible and therefore excluded, resulting in a final dataset of 3,410 users (Weverse_BTS_User_Dataset.xlsx).

In total, we collected 167,456 posts and 484,437 comments from these users. For each post or comment, the dataset includes a timestamp, text, user nickname, and user profile URL. We then filtered only English-language text and analyzed the remaining text using LIWC-22, resulting in the Weverse_BTS_LIWC_Dataset.xlsx.