README.md

metadata

license: mit
task_categories:
  - summarization
language:
  - en
size_categories:
  - n<1K
tags:
  - text

This dataset includes all recorded (spoken and written) addresses to the United States Congress from the President of the United States of America.

The addresses span all Presidents, but differ in their modality. From 1801 to 1913 the addresses were provided in writing to Congress with the remainder provided in person in the form of a speech.

Each address was scraped from The American Presidency Project, a website supported the UC Santa Barbara. Their mission statement is to, "be recognized as the authoritative, non-partisan on-line source for presidential public documents." A truly admirable goal.

The main index of the addresses are availble on this page.

The dataset is comprised of a single Parquet file with the following columns:

• potus : The full name of the President of the United States that gave the address
• date : The date the address was given.
• categories : The type of speech, either "Spoken" or "Written"
• speech_html : The unaltered or cleaned text of the speech scraped from the website.
• no-contractions : The speech with all contractions removed.
• tokens : An array of all words in the speech; all punctuation and special characters are now removed.
• tokens-nostop : An iteration of the tokens column; all stop words are now removed from the array.
• lemmatized : An iteration of the tokens-nostop column; this array contains a lemmatization of all the words from the previous column.