Educational Big Data and the Promise of Social Media

Educational Big Data and the Promise of Social Media

DOI: 10.4018/978-1-5225-5161-4.ch005
OnDemand PDF Download:
No Current Special Offers


In Chapter 4, the authors focused on some tools and mindsets that are beneficial for conducting analysis and research in a big data context. In this chapter, they turn their focus to the “data” part of big data and examine some interesting sources to begin to work with. Social Media and related digital communications are the most prominently featured exemplars, but they also discuss other sources of data that can be analyzed. Finally, the authors also survey some interesting recent research being done both in Community of Inquiry and elsewhere that highlights strong data analytics approaches, interesting data sources, and novel conceptualizations of big data-type questions.
Chapter Preview

The Richness And Multitude Of Ones And Zeros

The interesting data we will talk about is almost exclusively digital. Some of it is fundamentally new types of data, like clickstream and social media data. Some of it is merely a digital version of data that has always been around in one form or another, like large collections of student writing and other student-level information. We are focusing on digital data streams for a few reasons which make them a very appealing alternative to analog data:

  • 1.

    Availability: Digital data is generated in large quantities, is easy to store, and is often readily available. Many institutions or organizations likely already have large quantities of digital data, lessening the data collection and processing burden that would otherwise fall to the researcher;

  • 2.

    Richness: The ease of collection makes gathering rich datasets, not just large ones, feasible for a wide range of tasks;

  • 3.

    (Relative) Ease of Analysis: Since computers work natively with digital data, we can use computational methods (such as those in chapter 4) to churn through data much faster in digital form than in analog form; and

  • 4.

    Scalability: With a little care, techniques can be designed to scale up from a few hundred data points to a few hundred thousand, and entire pipelines--from data collection to cleaning to analysis--can be automated.

Complete Chapter List

Search this Book: