Citation

Song L, Li RYM, Yao Q. Safety Sci. 2022; 147: e105602.

Copyright

(Copyright © 2022, Elsevier Publishing)

DOI

10.1016/j.ssci.2021.105602

PMID

unavailable

Abstract

To study the social structure of English and French Tweets of occupational safety and their sentiment distributions, this study applied NodeXL and MeaningCloud to analyse 17,147 English Tweets and 16,618 French Tweets about "occupational safety" in Twitter. We found that French and English Twitter users who are interested in this topic did not usually interact. While top English Twitter influencers were professors, top French influencers were government officers and individuals. Clusters of Twitter members interested in occupational safety had a low tendency to reach people in other groups. Most failed to make good use of weak ties to increase their impact and shared information about occupational safety outside their circle of friends. This overthrows previous research that Twitter's social network was built based on the weak tie: Twitter users follow commentators, celebrities, and opinion leaders who do not know personally. Besides, we also conducted sentiment analysis via machine learning algorithms. We found that the more positive sentiment of an English Tweets, the more likely it will be retweeted. Yet, the more negative sentiment of a French Tweets, the more likely the Tweets will be retweeted. Thus, negative occupational safety Tweets have stronger impacts than positive ones among French but not English Tweets. While sentiment analysis results of French Tweets indicated that most Twitter users discussed occupational safety issues with a neutral tone, the number of extreme negative in French Tweets was a lot more than that of English. That reflects languaculture differences, and informal institutions impact users' behaviours.

Language: en

Keywords

New institutional economics; Occupational safety; Sentimental analysis; Social network analysis; Weak tie