| Summary: | Master of Science === Department of Agricultural Economics === Glynn T. Tonsor === Over the past decade, social media has risen from an emerging novelty to the normative form of expression for many Americans. As these platforms have risen in popularity, researchers have recognized the potential for capturing information users are self-reporting about their beliefs and preferences. Simultaneously, social media corporations have become privy to the value of this information being freely shared by consumers and have safeguarded much of their historical data to monetize the data. Faced with both an enticing new source of data, but a steep price to obtain it, researchers must evaluate the potential gains that can be extracted from the often difficult to analyze data.
This study explored the acquisition of social media, namely Twitter, data and the potential uses in the field of agriculture economics. A contract was secured with Sysomos, a social media analytics firm, in July of 2017 to collect raw Twitter data over the proceeding thirteen months. Changes in frequency of tweets and sentiment scoring of tweets were used to attempt to explain election results from November 2017 proposed legislations pertaining to marijuana and minimum wage as well as to explain and predict changes in the stock prices of selected publicly traded firms in the food producing sector. Twitter frequency changes were then compared to changes in traditional print media articles in an effort to determine the exchangeability of the two media sources when used to track events pertaining to animal health.
Results of this study suggested that Twitter data possess little power to explain the studied election results, but creation of a strong model was difficult due to the limited number of months of data available. Changes in the frequency of tweets were not found to be a strong indicator of changes in the stock market on the average day, but were shown to explain potentially highly valued information to investors on days with large changes in price. Twitter and traditional print media were shown to be unique sources of data when exploring the topic of animal health events.
|
|---|
