Employing Machine Learning for capturing COVID-19 consumer sentiments from six countries: A methodological illustration

  • By Kirti Sharma
    Associate Professor
    Bodo Schlegelmilch, Vienna University Of Economics And Business, Austria
    Sambbhav Garg, University Of Petroleum & Energy Studies, Uttarakhand

In this paper, the authors aimed to illustrate the scope and challenges of using computer-aided content analysis in international marketing to capture consumer sentiments about Covid-19 from multi-lingual tweets.  The study is based on some 35 million original Covid-19-related tweets. The study methodology illustrates the use of supervised machine learning and artificial neural network techniques to conduct extensive information extraction. The authors identified more than two million tweets from six countries and categorised them into PESTEL (i.e. Political, Economic, Social, Technological, Environmental and Legal) dimensions. The extracted consumer sentiments and associated emotions show substantial differences across countries. Our analyses highlight opportunities and challenges inherent in using multi-lingual online sentiment analysis in international marketing. Based on these insights, several future research directions are proposed. First, the authors contribute to methodology development in international marketing by providing a “use-case” for computer-aided text mining in a multi-lingual context. Second, the authors add to the knowledge on differences in Covid-19-related consumer sentiments in different countries. Third, the authors provide avenues for future research on the analysis of unstructured multi-media posts.