Emotional Markers As Indicators of Investor Attitudes: EDA Sub-process Proposal

The goal of this study is to suggest a method for generating time series from social media by using emotional markers. Emotional markers are defined as words (n-gram strings) and are identified based on their frequency. This simple method can be used to support the results of traditional positive/negative sentiment analysis or to create a new research strategy. The primary idea is to construct data subset depending on the domain knowledge and the decision-making psychology. We regard the recommended steps to be exploratory data analysis actions and to be conducted in the time series creating module. We used March 2020 as an example because financial markets experienced large declines due to the COVID-19 pandemic. As a result, the interpretation is more customized and intuitive than it would have been otherwise in other research processes.