Dynamics of emotion trends in Canadian Twitter users during COVID-19 confinement in relation to caseloads: Artificial intelligence-based emotion detection approach
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Multiple waves of COVID-19 have significantly impacted the emotional well-being of all, but many were subject to additional risks associated with forced regulations. The objective of this research was to assess the immediate emotional impact, expressed by Canadian Twitter users, and to estimate the linear relationship, with the vicissitudes of COVID caseloads, using ARIMA time-series regression. We developed two Artificial Intelligence-based algorithms to extract tweets using 18 semantic terms related to social confinement and locked down and then geocoded them to tag Canadian provinces. Tweets (n = 64,732) were classified as positive, negative, and neutral sentiments using a word-based Emotion Lexicon. Our results indicated: that Tweeters were expressing a higher daily percentage of negative sentiments representing, negative anticipation (30.1%), fear (28.1%), and anger (25.3%), than positive sentiments comprising positive anticipation (43.7%), trust (41.4%), and joy (14.9%), and neutral sentiments with mostly no emotions, when hash-tagged social confinement and locked down. In most provinces, negative sentiments took on average two to three days after caseloads increase to emerge, whereas positive sentiments took a slightly longer period of six to seven days to submerge. As daily caseloads increase, negative sentiment percentage increases in Manitoba (by 68% for 100 caseloads increase) and Atlantic Canada (by 89% with 100 caseloads increase) in wave 1(with 30% variations explained), while other provinces showed resilience. The opposite was noted in the positive sentiments. The daily percentage of emotional expression variations explained by daily caseloads in wave one were 30% for negative, 42% for neutral, and 2.1% for positive indicating that the emotional impact is multifactorial. These provincial-level impact differences with varying latency periods should be considered when planning geographically targeted, time-sensitive, confinement-related psychological health promotion efforts. Artificial Intelligence-based Geo-coded sentiment analysis of Twitter data opens possibilities for targeted rapid emotion sentiment detection opportunities.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it