Detecting Community Depression Dynamics Due to COVID-19 Pandemic in Australia

Area of Science:

• Public Health
• Computational Social Science
• Mental Health Research

Background:

• The COVID-19 pandemic has led to a global increase in mental health issues, with depression being a major public health concern.
• Depression poses significant personal and societal costs, impacting emotional, behavioral, and physical health.
• Understanding depression dynamics is crucial for addressing the mental health crisis exacerbated by the pandemic.

Purpose of the Study:

• To investigate community depression dynamics influenced by the COVID-19 pandemic using Twitter data.
• To develop and apply a novel classification model for identifying depression cues in social media content.
• To analyze the impact of COVID-19 and related events, including government measures, on depression levels.

Main Methods:

• Utilized user-generated content from Twitter, specifically focusing on users in New South Wales, Australia.
• Developed a novel approach combining multimodal features (emotion, topic, domain-specific) with Term Frequency-Inverse Document Frequency (TF-IDF).
• Built depression classification models to extract depression polarities from tweets related to the COVID-19 period.

Main Results:

• The study found a notable increase in depression levels among Twitter users after the onset of the COVID-19 pandemic.
• Government-imposed measures, such as state lockdowns, were correlated with heightened levels of depression.
• The proposed model effectively captured depression cues and their fluctuations during the pandemic.

Conclusions:

• The COVID-19 pandemic has had a discernible negative impact on community mental health, specifically increasing depression.
• Social media analysis provides valuable insights into public mental health trends during health crises.
• Policy interventions during pandemics require careful consideration of their potential mental health consequences.