Online discussion questions – Day 34 (Apr 8) – Beyond Numbers
- Any examples of sarcastic tweets or reviews for a company that is not straightforward to interpret as positive or negative?
- Any other examples of analysis of unstructured data?
- Any data source for sentiment analyses other than Twitter or Yelp?
15 Responses to Online discussion questions – Day 34 (Apr 8) – Beyond Numbers
Leave a Reply
You must be logged in to post a comment.
The tweet “@comcast tell thesa im sorry for wasting her time, maybe i can take her out some time to make it up to her…” can be interpreted as a sarcastic comment meaning that the customer had to cancel plans due to dealing with Comcast, however it can also been seen the other way around, the customer wasted someone as Comcast’s time and wanted to apologize. It is unclear whether this tweet is sincere or sarcastic.
On yelp, a female customer left a comment for Chick-fil-A stating “My lover Karen and I adore this place so much we’re considering holding our commitment ceremony here. Hope they do weddings!” Naturally, nothing seems to be wrong about the comment. In fact, it seems positive. However, if the viewer of the comment knew about the Chick-fil-A same sex marriage controversy, the viewer would understand the sarcastic tone of the comment.
An example of a sarcastic tweet could be: “I mean, seriously, having only two choices for cable and Internet is AWESOME. Especially when @Comcast is oh so reliable!” – this cannot be interpreted as being either positive or negative because while it uses words that have positive connotations (awesome), it is also exaggerated with specific words (seriously, oh so) which may make the audience believe that it is sarcasm.
An example of a sarcastic tweet is: “I heard McDonald’s chicken’s are grass feed and are naturally raised. SO HEALTHY, SO PROUD go McDonald’s!!!!” If someone looked at this tweet they would think that this person is praising McDonald’s. They would see key words such as ‘proud.’ However, someone who knows the background of the production of McDonald’s chicken would know that this person is mocking the company. Other data sources for sentiment analyses could be Youtube comments or product reviews on Amazon.
An example of a sarcastic tweet is ” I’m so glad that Taco Bell is no longer the cheapest place to buy gas.” This is sarcastic because when they say gas they don’t mean gasoline but flatulence which people can get at other restaurants now that are cheaper.
An example of a sarcastic comment from YELP is : “I really wish there was a chipotle near where I lived. I would probably go everyday. Hmm, maybe it’s a good thing I don’t live near one.” It is sarcastic because the person is somewhat contracting himself, so it is hard to tell if the comment is positive or negative.
Other sources of for sentiment analysis may be Youtube comments and Facebook posts/comments.
I found an example of a sarcastic tweet for American Airline, which says “Ur flight attendants are not “allowed” to help me lift up my carry on to the overhead bin? What if you have issues lifting it? @AmericanAir.” This tweet sounds negative to point out the unkindness of American Airlines’ flight attendants. Interesting is, the person who left this comment used the word, “lift up” to lift her complaints on Twitter in terms of the flight attendants’ not lifting up her baggage on the overhead cabin.
There was a yelp review for a McDonalds in New York: “The other good thing is the lines. There are really no lines, but rather a crowd. So you just push past people to the front. Something the tourists arent used to, but gives the lokes an advantage.” It uses a positive(“good”), but then goes ahead and describes a seemingly negative situation, so it’s hard to tell right away if this is positive or negative overall.
I found a comment on a yelp review for a fast food restaurant saying, “I looooooove finding questionable chunks of meat in my chicken nuggets…” Although this customer is using the words “love” they are dragging in out in a way that communicates sarcasm. A lot of people may glance and find the positive words without using context clues to determine the tone of the person who said it even though it may be a comment like this one on yelp where you are just reading the review rather than hearing it. Many times even though positive words are being used, the context in which they are used can be negative.
An example of a misleading comment on Yelp was left for Wendy’s saying:
Chicken nuggetz be crispy like you never seen. I tried one and I was like WHHAAT. Are you serious?
This type of review is very difficult for a computer to interpret because it showcases both positive and negative features.
An example of a sarcastic tweet is : Alyssa Lacey @AlyssaLaceyx 18m18 minutes ago
Hate to admit it but sheetz is actually so much better than wawa
It is not a very straightforward tweet, because it has the word Hate in it.
Another source for a sentiment analysis would be YouTube or Facebook.
An example of a misleading or sarcastic would be:
@BrennenCTaylor: “Chipotle sounds dank af rn”
This tweets is misleading because the commenter used a slang terms that could be either positive or negative depending on the context or the connotation the words are used in.
An example of a saractic tweet of Startbucks is:
@Xiwei Chen: Thanks Starbucks the staff misspelling my name again
It is not very straightforward because it start with “Thanks”.
The reviews on the “Rate my professor” are other examples of unstructured data. We can also do sentiment analysis on them. Through the analysis, we will know the actual students’ views of their professors.
@KendallThePlug i hate you. this one was an actuall mac … brand new, i opened it up and got it stolen the next day
Even though she was trying to indicate how she hated that her mac was stolen, the program would characterize this as a negative tweet.
“What a disgrace, there’s a woman crying at the @Ryanair check in desk who’s been made to pay more for emotional baggage”
he doesnt really mean that a baggage is emotional but he is trying to say that, when a women cry. he is trying to say that women cry alot and sometimes over nothing when there emotions play into effect.