This week we started out with python. We were asked to make a UNIX command-like python script for text manipulation. In the previous assignment I struggled with line length and that kind of came up again. I played around with the .replace() method and escape characters. I thought that by inserting “\n” into a line it would then be considered a new line in the rest of the script but it did not seem to work that way. Then I decided to try to remove a specific list of words (in this case conjunctions) and I was able to get that working in a for-loop. Eventually I did something similar to the two columns of words I made last week.
This time I chose a Federal Reserve statement on interest rates and a CNN news release about that release as the two source texts. I was able to go through each article and pick out all of the capitalized words over 3 characters in length and print them as a list. Then I added a randomized conjunction to each and created two columns with mostly proper nouns. The output reads fairly well and does seem to juxtapose the institutional and authoritative way the fed speaks with the more speculative and popular language of the CNN article.
Here’s the python code:
Here’s the output:
Information and Federal so
Federal nor Reserve and
Open so Trump so
Market so America’s for
Committee so Wednesday or
December nor President or
Household nor Trump’s or
Measures nor Trump’s for
Inflation so U.S. so
Committee’s yet That but
Market-based but Faster so
Consistent so Right but
Committee or Americans but
Committee for Trump and
Near-term for Republicans nor
Committee or Fed’s but
Committee and Trump nor
Committee so Mexican but
This nor Trade nor
Committee so Mexico so
Committee or House for
However, and Republicans nor
Committee for Supporters but
Treasury nor Some yet
This but Trump’s yet
Committee’s so Congress. nor
Voting for Michael or
FOMC or Arone, for
Janet nor State for
Yellen, but Street so
Chair; but Global but
William for Advisors. nor
Dudley, and Trump’s yet
Vice or America’s yet
Chairman; yet Francisco and
Lael but President or
Brainard; but John and
Charles but Williams but
Evans; or Financial but
Stanley so Times and
Fischer; or Jan. for
Patrick yet March or
Harker; yet Investors or
Robert for March, but
Kaplan; yet Group. or
Neel for
Kashkari; for
Jerome for
Powell; and
Daniel or
Tarullo. nor