It turns out 20,000 review is a lot, and it’s not all smooth sailing. We began running into difficulty in separating data into meaningful categories (especially since 1/3 of people didn’t write more than half a sentence), and our code began to have some issues. Besides looking for solutions, we did some more detailed analysis on the smaller sample sizes, and gave our report.
Hashmap diagram (what we are trying to get to work):
