This is my textual data visualization project for the Data Vis & Info Aesthetics class during my study at the MS Data Visualization program at Parsons School of Design.
I analyzed the past 10 years' State of the Union addresses to Congress and compared each speech with the Declaration of Independence, most particularly, how the keywords/American values of "American", "Equality", "Life", "Liberty" and "Happiness" were reflected in these speeches.
The American Promise
Data Storytelling
I pre-processed the corpus using normalization and tokenization methods and then I used the Gensim Word2Vec model to train and output top most similar words related to the forementioned keywards in each speech. I then manually tagged and categorized the similar words into seven major themes such as "economy", "humanity" or "science & tech" for a more aggregated analysis. [Github Documentation]
Tools I used
- d3.js
- HTML5/CSS
- Python
- Word2Vec