Classifying Twitter Users with the Activity-Context Geography Model

I’ll be presenting a paper cowritten with my advisor, Dr. Yongmei Lu, at the International Conference on Location-Based Social Media in Athens, GA on March 13, 2015.

The paper is titled “Classifying Twiter Users with the Activity-Context-Geography Model.” Below is the abstract:

This paper reports on an examination of Twitter contributor data using the Activity-Context-Geography (ACG) Model. The ACG Model is developed to reflect Volunteered Geographic Information (VGI) data contributors’ characteristics that are related to their contribution of geographic data. The study examines the characteristics of tweet contributors along three aspects: Activity, Context, and Geography. The analysis found eleven clusters of tweeters by considering whether each contributor is a low-outlier, not an outlier, or high-outlier in each of the three aspects. As the data distribution can be heavily skewed by a few extremely large contributors, classifying Twitter contributors based on their outlier status can groups extreme contributors in one group rather than separately. Future work should examine how the different tweeter groups are separated in their usage of social media and their motivation for tweeting, and how their association or affiliation with places and locations may be reflected through social media.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.