Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Tweet properly: analyzing deleted tweets to understand and identify regrettable ones
Zhou L., Wang W., Chen K.  WWW 2016 (Proceedings of the 25th International World Wide Web Conference, Montréal, Québec, Canada, Apr 11-15, 2016)603-612.2016.Type:Proceedings
Date Reviewed: Sep 8 2016

Whoever thought that a 140-character method of social media communication would warrant so much academic attention? Its use in my classroom was brought about by a desire to have students at the leading edge of all types of information technology forms. Thus, they are required to tweet about what is going on in their country of choice in a comparative politics course, and about a current woman leader in a women and politics course. They never know when a future employer will require such skills, as indicated by the move of some companies to have applicants tweet their resumes!

This paper goes out of its way to take a look at a specific set of tweets: those that have been deleted because the tweeter regretted the contents of what had been transmitted to the public. The authors take pains to describe the process and methods used to arrive at their analysis and conclusions. They indicate the number of tweets analyzed (more than one million to start out with) for one week, which is whittled down to 4,000 by the time various factors are taken into account. They describe the painstaking method of culling from the millions of those that qualify as regrettable and why those that were rejected were done so beyond grammatical and typographical errors, rephrasing, and spamming reasons.

They focus attention on normal individual users rather than varied users, which they define as celebrities and organizations that normally publish a lot of tweets (the Kim Kardashians of the world) and have a tendency to delete more tweets than normal users.

The authors admit that whether an individual will regret a tweet is purely of personal choice and would be subject to too many factors that would make it difficult to train a mathematical computer model that could identify regrettable tweets. They do know that sensitive and private information having to do with offensive comments, alcohol/illegal drug use, and sexual activity is the type of material that often leads to regrettable tweets. The development of classifiers for these types of tweets is helpful, and required the work of three annotators that agreed on ten classifying categories most of the time: negative sentiment, cursing, sex, alcohol, drug, violence, health, race and religion, job, and relationship.

As a Twitter user, I’m fascinated by what the authors have tried to do. However, I have my doubts regarding its usefulness beyond academia.

Reviewer:  Cecilia G. Manrique Review #: CR144741 (1612-0912)
Bookmark and Share
  Featured Reviewer  
 
Social Networking (H.3.4 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Social Networking": Date
Social computing and virtual communities
Zaphiris P., Ang C., Chapman & Hall/CRC, Boca Raton, FL, 2009.  303, Type: Book (978-1-420090-42-0)
Sep 13 2011
A visual analytics approach to dynamic social networks
Federico P., Aigner W., Miksch S., Windhager F., Zenk L.  i-KNOW 2011 (Proceedings of the 11th International Conference on Knowledge Management and Knowledge Technologies, Graz, Austria, Sep 7-9, 2011)1-8, 2011. Type: Proceedings
Nov 4 2011
Secure collaborative social networks
Zhan J. IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews 40(6): 682-689, 2010. Type: Article
Feb 8 2012
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy