Social Pulse: How is the World Reacting to the COVID-19 Pandemic?

March 20, 2020
As the COVID-19 pandemic dominates all conversation, our desire to understand how people interact in times of crisis, led us, Clarabridge, to seek more information. How are people actually feeling about this global pandemic? In an effort to capture conversations and emotions about the economy, the virus, and a potential “new normal”, Clarabridge is gathering a constant stream of Twitter data for analysis. Throughout this series, we are sharing our insights—call it a “pulse check” on our society. Are we becoming more connected in a time of disconnect? Are we hopeful or truly feeling blue? And how does public sentiment affect business and the economy in the short and long term?
The Findings
Clarabridge leverages Natural Language Processing (NLP) and machine-learning augmented enrichments (Natural Language Understanding) to understand topics of conversation and derive additional insight on the “current state” of societal discourse. Our metrics summarize aggregated public Twitter data to show the overall sentiment and emotional intensity expressed within the data set.
Here are a few of our “fast stats” we’ll continue to update and share:
SOCIAL SENTIMENT |
-.25![]() About the Scale: -5 to +5 Clarabridge’s sentiment score shows how positively or negatively people are talking about their experience. |
EMOTIONAL INTENSITY |
2.23![]() About the Scale: 1 to 5 Clarabridge’s Emotional Intensity metric provides insight into how intensely a person is expressing emotion. |
Leveraging our existing understanding of language and emotions, we defined six categories of emotions: fear and confusion, anger and frustration, kindness and hope, joy and humor, anticipation, and sadness. These emotions trend highest across our sample.
Through Natural Language Understanding (NLU) enrichments, Clarabridge assesses emotional intensity (a mild expression of fear such as “afraid” versus an intense expression such as “absolutely terrified”, for example). For the six categories of emotion shown above, we include here the strength of expressed emotion.
Society is afraid and confused and expressing it in full force through social channels. Individuals are frustrated and livid. Perceptions of the methods chosen by leaders to deal with the crisis show frustration on both sides of the aisle and the economic impact of the crisis leave many Americans worrying about financial instability. But aside from the known factors we see in the news, frustration comes from our inability to see the value in social good. For the first time in public data comes the debate society has always had: community good versus survival of the fittest. High frustration tweets include members of vulnerable populations stating things like:
On the contrary we see several messages of kindness also expressed in the forum. Individuals express passionate appreciation for humanity including:
And while we’re seeing humor on a smaller scale, some of these jokes share a sense of joy:
These anecdotes show us that we are collectively embracing the small positive actions that flood the world wide web.
What’s Next
Nicole Martin is currently a consultant at Clarabridge. Prior to Clarabridge, Nicole received her Master of Public Health in Epidemiology from The George Washington University. During her time at GW, Nicole wrote her graduate paper on sexuality, sexual behavior, and mental health. In addition, Nicole taught as a Graduate Assistant for the Biostatistics Department at The George Washington University. During her time at Clarabridge, Nicole has worked with healthcare accounts to enrich their analytic capabilities, created customer journey maps for property and casualty insurance companies, and continued to support innovation for clients across various industry verticals.

About The Social Pulse Series
Clarabridge has embarked on an independent research project to actively analyze the “emotional pulse” of social media users worldwide during the COVID-19 pandemic. The effort’s main goal is to assess how people are feeling using Clarabridge’s Natural Language Understanding to glean insights from millions of unstructured data records. We hope to inform the public, provide insights to the scientific community and educate Clarabridge customers. The analyses in this series leverages Twitter data collected beginning March 12th using keywords such as “coronavirus,” “covid19”, and “covid-19” from Twitter. We continue to refine data collection and models as the situation evolves.
Helpful Resources:
Questions about Coronavirus? Check out the following resources:
Centers for Disease Control and Prevention (CDC)
World Health Organization (WHO)
Directory of Local Health Departments