Analysis of social media sentiment and economic indicators during the Covid-19 pandemic: a preliminary investigation

We analyse the sentiment of New Zealand-based Twitter data for two and a half years around the Covid-19 pandemic. We find an increase in the ratio of negative sentiment, particularly at the onset of the first nationwide lockdown in March 2020, the first Auckland-only lockdown in August 2020, and the arrival of the Omicron wave in late 2022. We then correlate the New Zealand electronic card transaction data with our sentiment analysis results and the ANZ Roy Morgan Consumer Confidence Rating. We find that these two sets of correlations are surprisingly similar in a few key categories. We hope to use this paper to promote the use of sentiment analysis on a wide range of texts; one promising application would be a real-time measurement of consumer confidence.