CMStatistics 2017: Start Registration
View Submission - CMStatistics
Title: Complexity of United State of America presidential speeches Authors:  Valerio Ficcadenti - University of Macerata (Italy) [presenting]
Roy Cerqueti - University of Macerata (Italy)
Abstract: The aim is to explore the rhetoric dynamics of a large collection of United States of America Presidents' speeches. In particular, speeches are viewed as complex systems and are analyzed through rank-size laws, being the words of each speech ranked in terms of their frequencies. At this aim, a best-fit procedure with Zipf or with Zipf-Mandelbrot laws is performed over the 951 talks individually. Thanks to these estimations, it is possible to reach interesting conclusions on how 45 United States Presidents, from April 30, 1789 (Washington) until February 28, 2017 (Trump), have delivered political messages. Our analysis shows some remarkable regularities, not only inside a given speech, but also between different speeches. We discuss the political and linguistics aspects. The building of the dataset itself represents a relevant step of the study. In this respect, by using a web scraping routing on the Miller Center website, a large span of 978 speeches have been downloaded. After a pre-processing phase, the set is reduced to 951; for each one, the words' frequencies are stored and the analysis is performed.