Issues for nltk_guidelines: NLTK Methods - End

[rachelrakov]

• In section "NLTK Methods with the NLTK Corpus", when importing matplotlib, mention what it is - define it.

• In section "Searching for Words":

  • Typo: "What about Sense and Sensibility?" - add question mark
    *Define "phatic function", as compared to a semantic one; may not be known by all

• In section "Positioning Words"

  • Define what a pipeline is

• In section "Types vs Tokens":

  • Typo: "we would need a count of the 'lol's." is missing and end parenthesis
  • You bring in list comprehensions here with the code text1_tokens= [t.lower() for t in text1 if t.isalpha()] - I am not certain if participants will have ever seen a list comprehension before (not sure if Patrick covers it in the Python workshop) - take a minute to explain them.
  • Might be worth mentioning that .isalpha() ignores punctuation - that is not necessarily intuitive
  • Restore non-code after the above example back to words in markdown format - for awhile it's all in code format, which makes it very difficult to read

• In section "Cleaning the Corpus":

  • Might be work explaining how len(set(text1_tokens)) shows how many unique words there are - unless you have predefined set, which I did not see.
  • It is important, if you are billing list comprehensions as a computationally faster way of completing what can be done in lists, to define them and mention that they are computationally faster as early on as you can.
  • Typo: "but just for illustration, we can try it out with a instead (Porter is the most common)." Need a word between a and instead.
  • I am not certain that Patrick will have taught dictionaries in his lesson - check with him before you mention dictionaries in trying to describe Frequency Distributions.
  • I am also not certain if Patrick will have taught pass during his workshop - might need to explain what that does as you are going through the code.

• In section "Make Your Own Corpus":

  • Once again, make the text that isn't code here back into a markdown format, for ease of reading (might just be a matter of making the cell a markdown cell, if you wrote this in jupyter notebook)

• In section "Part of Speech Tagging":

  • Typo - "I'll make an epty dictionary to hold my results" - should be "empty"
  • Once again, you may need to explain the syntax of creating a dictionary here, if that is not covered in the Python tutorial, as well as what a dictionary is and how it works. Check with Patrick.
  • Fix the formatting at the end - once again, it's all in code format, making it hard to read. Everything that isn't code should be in markdown cells.

• Are you going to add what is at the end of your jupyter notebook for this session into your nltk_guidelines markdown? There is discrepancy there; not sure if that is deliberate or not.