Spatial Analysis of Social Media’s Proxies for Human Emotion and Cognition

Apparent and latent knowledge claims made by social media authors are uncovered via Natural Language Processing tools and techniques. This phenomenon presents several fundamental issues as researchers distinguish correlation factors between an author’s endogenous self-described and self-validated data disclosure and exogenous relationships. While social media processing protocols produce near-instantaneous analysis of a social dataset, studies examining such tools often overlook the strength of the correlation relationship. In addition, they frequently neglect to report an assessment of the normality of data distribution, disregard the rationale for choosing Pearson’s or Spearman’s test of correlations, and sometimes even use contrasting interpretations of correlation coefficients. This study proffers a correlation study research approach via direct inquiry that strengthens the theoretical foundation for identifying the relationship among social media emotion, sentiment, and cognition across diverse locations concerning real-world events. It attempts to answer the main research question: Does any relationship exist between social media’s sentiment and cognition variables? It constructs a robust research-based analysis to validate the main question and fill the gap that other projects fail to address. Furthermore, it presents and discusses methods for investigating and interpreting correlation matrices and scatterplots. Moreover, it answers the research question and attempts to prove spatial relationships among variables. Last, it guides and describes future work of a predictive artifact that will input, process, and visualize a spatiotemporal, NLP processed, social media dataset and its integration with Pearson’s and Spearman’s correlations, and visual data constructs.