Article Preview
Top1. Introduction
The emergence of Web 2.0 in 2000’s (O’Reilly, 2005) has revolutionised the way people interact through the internet encouraging the information sharing, participation, and collaboration of individuals to create and consume information through blogs, wikis, social networking sites, etc. (Rennie and Morrison, 2013). The advancement in technology has also transformed education with reducing the time and space constraints associated with traditional learning methods. Massive open online courses (MOOCs) are gaining popularity for providing online courses to a larger mass for free or with very nominal costs. For instance, the MOOC market is expected to grow from 1.83 billion USD in 2015 to 8.5 billion USD in 2020 (MarketsandMarkets, 2015). Despite the advantages of asynchronous learning of MOOCs and being offered at almost free of cost, the attrition rate is very high (Clow, 2013; Perna et al., 2014). Therefore, it is essential to understand the users’ attitude towards MOOCs, and social media is the best place to understand crowd sentiments considering the massive amount of data being created.
Text analysis from social network sites has helped researchers to understand the research problems and provide suggestions in various areas including product sales (Chong et al., 2016), predicting election results (Rill et al., 2014), comprehending public sentiments in a sensitive situation like disasters (Bai et al., 2016), etc. Similarly, the MOOC-related public sentiments can be mined to understand the opinion of the crowd as MOOC users actively use social media (Delgado-Kloos and Munoz-Organero, 2014; Kravvaris et al., 2016). Twitter can be used as a competitive intelligence source to discover the customers' sentiments about a brand and its competitors in real-time (Jansen et al., 2009). The high speed (140 characters), open community, and asymmetric nature of Twitter allows the tweets to be visible by a wider user base and makes it easier to capture the dynamics of interaction among users, unlike other available social communities. Since MOOC users actively use Twitter to share the topics, question-answers, etc., their sentiments can be captured, and the public opinions regarding MOOCs can be mined.
Many researchers (Chen, 2014; Enriquez-Gibson, 2014; Shen and Kuo, 2015; Rui et al., 2013; Wen et al., 2014; Xia, 2014) have performed text mining on MOOCs from social media data, but they are mostly based on generic MOOC-related keywords or one-or-two MOOC providers. There is little research from the perspective of both MOOC users and providers such as, how MOOC users perceive and engage in social media and how MOOC providers keep their users engaged. The present study aims at seven major MOOC providers namely, Codecademy, Coursera, edX, Khan Academy, Skillshare, Futurelearn, Udacity, and examines (1) the temporal trends, i.e., daily, weekly, and monthly trends of MOOC providers’ related data from Twitter, (2) Association rule mining of tweets to find interesting patterns, (3) Sentiment mining of tweets and categorizing them into positive and negative opinions, (4) Influencer analysis, and (5) Engagement analysis from the provider side.
This paper is structured as follows: Section 2 presents the literature review; Section 3 gives the details of the research framework. Section 4 presents results obtained followed by the discussion in Section 5. Lastly, Section 6 presents the conclusion of the study, limitations and future scope of research.