Comparing US Presidential Speeches.
With Barack Obama and Donald Trump two presidents very different from one another led the United States of America back-to-back in the last decade. Barack Obama was the first black president in the history of the nation and introduced progressive policies like the American Health Care Act during his two terms in office. Donald Trump on the other hand followed very conservative and nationalist politics with harsh immigration policies propagating the motto “America First”. The popular consensus is that there are stark differences in their corpora. I want to find out if there is objective support for these felt differences.
In two previous papers I laid the groundwork for comparing the texts of these two presidents, creating corpora of their public speeches.
The goal of this thesis is to use a multitude of digital methods to show differences between the two corpora. These methods include simple digital analyses like type-token-ratio, word and sentence length, part of speech tagging and word frequencies, and more complex methods like readability scoring, topic modeling, sentiment analysis and authorship attribution with machine learning. This thesis will not include the political or societal impacts and interpretations of the data based results. The main emphasis will hereby lie on the methodology and generating data.
Code auf Github