Analysis of Public Sentiment Impact on Stock Market Indicators

This paper is devoted to the analysis of social mood influence on stock market indicators. In this research, public mood is determined through estimation of social network Twitter posts’ sentiment, whereas financial market indicators are manifested in the form of S&P500 and FTSE100 – largest American and British stock indexes. The analysis is separated into three parts. Firstly, we investigate relevant literature on previous research on public mood and stock market interconnections. Then, we introduce our own methodology that covers data retrieval, filtration and analysis (Natural Language Processing) techniques. Finally, we institute modelling process that examines dynamic time series associations through cross-correlation, causation and forecasting analyses that introduce Distributed Lag and Vector Autoregression statistical models. A lot of researchers previously attempted to investigate public opinion – financial markets relationships, but a very few of them outlined significance of opinions’ popularity in determining effects on stock markets. None of the papers cover this issue in detail, and we aim to amend the existing literature by adding popularity measurement techniques. This research covers highly volatile COVID-19 period, ranging from January to July, 2020. Thereby, besides analyzing impact of financial tweets, we aim to reveal effect that general (COVID-19) tweets impose on stock markets. This effect was not studied previously, so the results are novel. This research may be either applied in price prediction purposes or viewed as a basis for further theoretical analysis. Thereby, we infer that results of this paper may be widely applied both in academic and practical fields.