Thai Social Media Corpus for Depression and Anxiety Detection

This corpus is part of my PhD research, “Exploring Emotion Timeline
Patterns in Social Media for Automatic Identification of Depression and
Anxiety.” It contains chronological tweets from Thai users, with each user
tagged with depression and anxiety labels. The corpus also includes
emotion-intensity labels across 26 fine-grained categories for each tweet,
annotated manually by human experts. The corpus contains two files: one
providing user information and the other containing users’ chronological
tweets. The user information file (users_anonym.csv) provides depression
and anxiety labels for each user. For both labels, a value of 0 indicates
the absence of depression or anxiety, while a value of 1 indicates the
presence of the condition. The tweets file (timelines_anonym.csv) contains
the tweeting timelines of each user. Each tweet is tagged with a user ID,
timestamp, and emotion-intensity labels. To protect users’ privacy, all
sensitive information has been replaced with pseudonyms. This includes
names, addresses, dates, places of work or study, job titles, salaries,
telephone numbers, fax numbers, account numbers, certificate or licence
numbers, email addresses, social media profile names (or handles), user
mentions, website URLs, IP addresses, vehicle identifiers and serial
numbers, and other personal details. Description