presentation researcher mental health
The presentation, authored by members of the Researcher Mental Health Observatory (ReMO) and affiliated with the COST Action CA19117, outlines how Twitter can serve both as a community platform and a data source for studying academic mental health. It begins by stating four objectives: exploring mental‑health conversations on Twitter, reviewing literature linking mental health and Twitter, familiarising the audience with Twitter‑based research methods, and highlighting ethical concerns. The slides showcase a range of academic mental‑health Twitter accounts and hashtags, emphasizing Twitter’s reach (over 313 million monthly users) and its therapeutic potential. Methodologically, the team describes using the Twitter Academic API (free, up to 10 million tweets/month) combined with text‑mining techniques—sentiment polarity scoring, single‑word and n‑gram analysis, LDA topic modelling, and network visualisations. Quantitative results include 36 867 unique tweets from 28 954 users, with geographic breakdowns (USA, GB, Australia, South Africa) and temporal trends from 2017 to 2022. Sentiment analysis reveals a near‑even split between negative (48 %) and positive (45 %) tweets, with country‑specific polarity trends. Keyword and single‑word polarity tables identify words that drive sentiment (e.g., “well‑being” positive, “deadline” negative). N‑gram tables expose recurring positive and negative phrase patterns across countries. Network graphs map hashtag co‑occurrences and community structures, while LDA uncovers thematic topics and their average polarity. Two case studies illustrate (1) keyword/hashtag discovery for mental‑health discourse and (2) profiling specific Twitter communities (e.g., Academic Chatter, PhD_Genie). The presentation concludes with ethical guidance (public vs private data, committee approval), a call for more qualitative studies, practical advice for leveraging Twitter to build supportive communities, and information about ReMO’s network resources and the 2022 Budapest conference.
Topics
Mapping the Landscape of Academic Mental‑Health Communities on Twitter
Shows prominent accounts, hashtags, and follower counts, illustrating the existing ecosystem of support and advocacy among scholars.
Methodological Framework: API Access, Text Mining, and Analytic Techniques
Details the use of the Twitter Academic API, data extraction limits, and analytical pipelines (sentiment scoring, keyword extraction, n‑grams, LDA, network visualisation).
Quantitative Overview: Tweet Volume, Geographic Distribution, and Temporal Trends
Presents counts of unique tweets and users, country‑level breakdowns (USA, GB, Australia, South Africa), and quarterly tweet trends from 2017‑2022.
Sentiment Polarity Analysis Across Countries, Years, and Quarters
Reports percentages of negative, positive, and neutral tweets, and shows how average polarity shifts over time and differs by nation.
Keyword and Single‑Word Polarity Insights
Lists high‑impact words with associated polarity scores, highlighting terms that drive positive (e.g., “well‑being”) versus negative (e.g., “deadline”) sentiment.
Network Visualisations of Hashtag Co‑occurrence and Community Structure
Shows graph representations of hashtag clusters (e.g., Academic Chatter, PhD_Voice) and explains how they map discourse and support networks.
Case Study 1: Keyword and Hashtag Discovery for Mental‑Health Tweets
Demonstrates how to generate word clouds, select hashtags, and use tools like Hashtagify to locate relevant conversations.
Case Study 2: Profiling Specific Academic Mental‑Health Twitter Communities
Analyzes tweet counts, word usage, and network graphs for communities such as Academic Chatter, Happy Researchers, and PhD_Genie.