Social Sciences

As much as the media covers natural disasters extensively, little is known about the underlying effects aside from the “who was hurt or killed and what was destroyed” coverage of the disaster. Natural disasters often cause geographic displacement of individuals from their homes and communities to new ones. These events result in more than just the loss of physical property, as community and neighborhood attributes encompass many different aspects of an individual’s social relationships and the cultural institutions. Moving to a new community changes these attributes and will affect an individual’s political attitudes and behaviors. The ongoing post-Harvey migration provides a unique opportunity to examine the effect of social contact and political context in a case where these encounters could not have otherwise been anticipated by the individuals affected. This research study seeks to examine the political consequences of Hurricane Harvey by studying post-Harvey migration patterns as an unexpected or exogenous shock via data analysis techniques such as Natural Language Processing using C++ programming. This analysis will allow us to study variations in the decision of displaced individuals to move to different areas and how this affects their associated experiences once they move, and how people’s political attitudes and behaviors change in response to rapid demographic shifts in their communities in the future. This research is therefore able to address “How does local context affect political opinions and behaviors?” Using hand-coded twitter data, we were able to create various predictive models coded in R, Python, and C++. My specific research focuses on building a C++ predictive model. The C++ uses a score based system on NLP and even caught mistakes made during manual coding. We were able to determine that natural disasters such as Hurricane Harvey do influence social and political behaviors of the people affected by using an analysis of social media such as twitter. We also determined that the C++ predictive model had a very accurate prediction on sample twitter data.