Rise and fall of the global conversation and shifting sentiments during the COVID-19 pandemic

Social media has been widely used for public health surveillance, and COVID-19 is the first pandemic in a highly connected online world. The authors analyzed 105+ million tweets (Mar 1–May 15, 2020) and Weibo posts (Jan 20–May 15, 2020) across six languages (English, Spanish, Arabic, French, Italian, Chinese), representing an estimated 2.4 billion citizens. They trained deep-learning classifiers to detect fine-grained emotions: positive (optimistic, thankful, empathetic), negative (pessimistic, anxious, sad, annoyed, denial), and joking. Conversation volumes showed a rapid rise and gradual decline in all six languages, with peaks driven by economic collapse and confinement measures (except Chinese, where confinement dominated). Emotional patterns were similar across languages: early surges mixed joking with anxious/pessimistic/annoyed, later shifting to increased positive states as the pandemic came under control, strongest in Arabic tweets.

Xiangliang Zhang, Qiang Yang, Somayah Albaradei, Xiaoting Lyu, Hind Alamro, Adil Salhi, Changsheng Ma, Manal Alshehri, Inji Ibrahim Jaber, Faroug Tifratene, Wei Wang, Takashi Gojobori, Carlos M. Duarte, Xin Gao