I have been intrigued by Twitter for sometime. When the Library of Congress announced in 2010 that it would archive public tweets some scoffed, but I cheered. Sure there is a great deal of inane tweeting going on out there, but also this is a huge resource that could be a boon for numerous researchers. With this massive amount of data surely some trends, ideas, and conclusions can be reached on many subjects. It is all a matter of figuring how to do so. This is why the article we read in class "Sentiment in Twitter Events," (Thewall, Kevan, Paltoglou, 2010) really piqued my interest. The authors devised a study of a month of English Twitter posts to assess whether "popular events are typically associated with increase in sentiment strength." They found strong evidence that important events in Twitter are associated with increases in average negative sentiment strength. This is just one study that deals with a small cross section of data.
I was curious to find what tools are available to assess the glut of Tweets. One that I found was Sentiment 140, which came out around the same time as this study. It can be used to "discover the sentiment of a brand, product, or topic. I search for "library" and it assessed Tweets for the last 3 minutes with library as a topic. The analysis for those 85 tweets were 55% negative and 45% positive. I could scroll through to read the Tweets that were analysed and I agreed with how it calculated positive v. negative. There are many, many more and Sentiment 140 provides a listing of many other similar sites.
So that might be interesting to see, but how can it be applied? I used to see how positively/negatively my employer (VCU) is viewed on Twitter. It provides an illuminating snapshot of user sentiment. However this type of tool can be employed on a much larger scale as was seen during the 2012 presidential election. The University of Southern California used Twitter to track in real time public sentiment leading up to the first presidential debate.
Not everyone is sold on the idea of tracking Twitter sentiment. In the article "Why the trick to analyzing Twitter data is more data," that came out during the elections last fall, there are several reasons why Twitter may not be very effective in gauging public sentiment. While it does indeed offer a vast amount of data there are drawbacks including:
Sources:
Bachelor, Lisa. "Complain on Twitter for an instant response," The Guardian, 12 May 2012.
Bialik, Carl. "Timing Twitter," WSJ 28 Sept 2012.
Gross, Doug. "Library of Congress digs into 170 billion tweets." CNN, 7 Jan 2013.
Harris, Derrick, "Why the trick to analyzing Twitter data is more data," Gigaom, 2 Oct 2012.
Sentiment 140, http://help.sentiment140.com/home
"Ahead of Debate, USC Project Uses Twitter to Measure Political Sentiment." SocialTech.com, 3 Oct 2012
I was curious to find what tools are available to assess the glut of Tweets. One that I found was Sentiment 140, which came out around the same time as this study. It can be used to "discover the sentiment of a brand, product, or topic. I search for "library" and it assessed Tweets for the last 3 minutes with library as a topic. The analysis for those 85 tweets were 55% negative and 45% positive. I could scroll through to read the Tweets that were analysed and I agreed with how it calculated positive v. negative. There are many, many more and Sentiment 140 provides a listing of many other similar sites.
So that might be interesting to see, but how can it be applied? I used to see how positively/negatively my employer (VCU) is viewed on Twitter. It provides an illuminating snapshot of user sentiment. However this type of tool can be employed on a much larger scale as was seen during the 2012 presidential election. The University of Southern California used Twitter to track in real time public sentiment leading up to the first presidential debate.
Not everyone is sold on the idea of tracking Twitter sentiment. In the article "Why the trick to analyzing Twitter data is more data," that came out during the elections last fall, there are several reasons why Twitter may not be very effective in gauging public sentiment. While it does indeed offer a vast amount of data there are drawbacks including:
- It's a large sample size, but still just a fraction of the population.
- Even among internet users, Twitter skews toward a younger, more-connected demographic.
- Twitters undercounts total tweets when the system can’t make a link between tweets and a given event.
- The number of tweets per minute on any topic will naturally rise as Twitter’s user count rises.
- Tweeting may be a sign that someone is less engaged in an activity (e.g., watching a presidential debate) than someone watching intently.
- Sentiment analysis can be skewed by who’s tweeting about an issue (e.g., ardent supporters only, adversaries only, or the public at large) From Carl Bialik, The Numbers Guy, WSJ
So while Twitter can be a great source of information on popular sentiment it cannot be used as the sole source of data. It must be analyzed in conjunction with other data sets in order to provide context.
The article does mention one area where Twitter might be a better tool for gauging sentiment and that is consumer satisfaction. A Twitter user can quickly offer up their unvarnished opinion on goods and services. This takes much less time than telephone surveys and focus groups to gauge opinion and consumers are often more honest about their experience when limited to 140 characters. Many companies have created Twitter feeds to address consumer complaints. I have vented my frustrations about Verizon on Twitter a few times. I did not get any satisfaction other than letting off some steam, but some consumers have found resolutions for their problems via a Twitter rant.
Sources:
Bachelor, Lisa. "Complain on Twitter for an instant response," The Guardian, 12 May 2012.
Bialik, Carl. "Timing Twitter," WSJ 28 Sept 2012.
Gross, Doug. "Library of Congress digs into 170 billion tweets." CNN, 7 Jan 2013.
Harris, Derrick, "Why the trick to analyzing Twitter data is more data," Gigaom, 2 Oct 2012.
Sentiment 140, http://help.sentiment140.com/home
"Ahead of Debate, USC Project Uses Twitter to Measure Political Sentiment." SocialTech.com, 3 Oct 2012
I have never really done much with twitter however, my students use it all the time. I really dont know how I feel about it-posting so much about ones self on the internet all the time??? I thought this article was interesting about the idea of twitter and the powers that run it. I thought you did a nice job with your ideas and points this week. I just think that something like twitter does give to many people the option to give their opinion on way to many things.
ReplyDeleteThanks for your comments. I agree that some people put way too much out there, but I have found Twitter to be a great way to connect with some people. I follow many archivists and librarians. People are always posting interesting links to articles that I would not have found otherwise. They also bring up some interesting questions from time to time. I have steered clear of those who post every two minutes about what they ate, random thoughts, etc. One of my pet peeves on Twitter is the inappropriate use of the hash tag. Tagging random words is not helpful nor are the huge run on tags that have inadvertent (albeit funny in some cases) consequences like the #nowthatcherisdead. Some people read this as "now that Cher is dead," or "now T(erri) Hatcher is dead," instead of "now Thatcher is dead."
ReplyDeleteStill while Twitter is not for everyone, I think it has its place. There is something to be said for making a concise statement in 140 characters without abbreviations. I am also excited about the possibilities of using Twitter data to study various concepts like sentiment. As with any data it is important to realize who is creating it and what its limitations are. That is why I appreciated Bialik's points.