Abstract:
|
Every year, the National Agricultural Statistics Service (NASS) of the U.S. Department of Agriculture (USDA) produces hundreds of reports, providing those in agriculture critical information. Since 2006, Twitter has become a viable mode in which millions of people disseminate and collect information. Since 2009, NASS has used Twitter as a means to highlight relevant information about the agency and information found within the many reports it publishes. As NASS and other agencies have become more adept at storing assorted types of metadata associated with their Twitter accounts, analytic programs, such as SAS, JMP, and R, have incorporated features that facilitate examining the dynamics involved when a person 'views' or reads a tweet. In this analysis, a replicable classification framework is applied to a sample of NASS tweets to evaluate what types of content elicit higher or lower viewership. In addition, descriptive statistics, text mining, and other data mining techniques are used to examine what factors are associated with the most views. The results of the analyses are discussed.
|