Reviewed 63 of 63 files at r1.
Review status: all files reviewed at latest revision, 11 unresolved discussions.

analyze.py, line 9 [r1] (raw file):
Shouldn't you be able to pull these names from the web scraped dictionaries?

analyze.py, line 20 [r1] (raw file):
This list comprehension is not really necessary.

analyze.py, line 43 [r1] (raw file):
Would be helpful to include comments for what parameters functions take.

analyze.py, line 65 [r1] (raw file):
Lot of similar variable names, a bit difficult to keep track of.

analyze.py, line 70 [r1] (raw file):
Looks good.

save.py, line 16 [r1] (raw file):
Good compartmentalization.

save.py, line 64 [r1] (raw file):
Would probably be useful to include inline comments explaining your regex, since you probably won't remember what it does when looking back.

save.py, line 97 [r1] (raw file):
For future reference, you could have used the built in function "update" on dictionaries to perform a merge. You also could have used an * parameter to take a list of any number of dictionaries to merge, rather than dictionaries and default.

writeup.txt, line 44 [r1] (raw file):
Probably would have made more sense to print the results with most frequent at the top.

writeup.txt, line 81 [r1] (raw file):
Not sure what you mean by probability of the frequency analysis. Frequency analysis should just be counting the number of times each person said each word.

Upon reading your code, I think you meant something about the word mainly being used by that candidate and not other candidates. Could definitely be explained better here.

downloads/data/.DS_Store, line 0 [r1] (raw file):
Make sure to include .DS_Store files in the .gitignore. They are some sort of OS X cache file and shouldn't be pushed to github.

Comments from the review on Reviewable.io