Food insufficiency and Twitter emotions during a pandemic

Area of Science:

• Social Sciences
• Public Health
• Computational Social Science

Background:

• The COVID-19 pandemic raised global concerns regarding food insecurity.
• Real-time analysis of social media, specifically tweets, offers a potential tool for food assistance providers to identify and address urgent needs.
• Understanding public sentiment during crises is crucial for effective resource allocation.

Purpose of the Study:

• To explore the sentiments and emotions expressed in food security-related tweets during the early stages of the COVID-19 pandemic in U.S. states.
• To investigate the correlation between these expressed emotions and contemporaneous food insufficiency rates.
• To assess the utility of natural language processing (NLP) for real-time food security monitoring.

Main Methods:

• Utilized natural language processing (NLP) techniques to analyze food security-related tweets from U.S. states early in the COVID-19 pandemic.
• Extracted dominant emotions and sentiments from the tweet data.
• Correlated extracted emotional data with food insufficiency rates reported by the U.S. Household Pulse Survey.

Main Results:

• Nationally, the emotion 'joy' was most prevalent in food security-related tweets.
• However, 'anger,' 'disgust,' and 'fear' showed statistically significant correlations with concurrent food insufficiency rates.
• State-level analysis revealed more nuanced correlations, including a negative correlation between 'joy' and food insufficiency.

Conclusions:

• Emotions expressed on Twitter, particularly negative ones like anger, disgust, and fear, can serve as indicators of food insecurity.
• NLP analysis of social media provides a valuable, real-time method for understanding public sentiment related to food security.
• These findings can inform targeted food assistance strategies and public health interventions during crises.