Managing and Visualizing Unstructured Big Data

Introduction

This essay expands on the notion of “Big Data” to open up alternative analytic opportunities, on certain components of the data, through a theoretical lens that is mobilized to offer an interpretation and visualization of the information contained in large amounts of data. It is useful first to examine the term “Big Data” in some detail. The term refers to a phenomenon which results from the fact that institutions and individuals are digitizing many different kinds of information leading to an exponential growth of the amount of data can that is being stored in the digital space. First, this expansion relates to the increase in data points as more records are added to the corpus of Big Data. Second, the idea of Big Data needs to be considered in terms of the details that are being digitized. The notion of Big Data should be considered both in terms of the breadth of the data in terms of number of data points (amount) and the depth of the data related to the various fields of information available for each record (details). Therefore, Big Data has become an object of analysis for a variety of groups, from academics to marketers, all of whom are interested in understanding how Big Data could provide highly granular voluminous information about people (see, e.g., Mitra, 2014c). Next, it is useful to examine the different categories of information that makes up “Big Data.”

Much of Big Data is numeric that is amenable to mathematical analysis. For instance, it is possible to easily count the number of tweets produced by an individual. Such counts offer the opportunity for companies such as Tweeter to offer information about what topics are popular at any moment in time. The segment of Big Data that offers the ease of analysis and visualization has been called “structured” Big Data. There is, however, another vast component of Big Data that does not allow for easy numeric analysis. This segment is made up of the utterances of the people who are self-generating the Big Data by voicing themselves in the digital space. An example of this segment of Big Data is the actual specific tweet produced by an individual or the specific photograph uploaded on photograph sharing spaces. In the case of the tweet, the language of the tweet contains information about attitudes and opinions, just as a photograph offers information about the individual who has captured the picture. This form of data requires a more nuanced and “qualitative” analytic process that would discover the intent of the authors and the meanings of the messages encapsulated in a microblog or picture. This segment of Big Data has been named “unstructured” Big Data. Not only is it difficult to analyze the unstructured Big Data but it is also difficult to visualize the findings of analysis. The qualitative process does not typically produce convenient charts and graphs. The analysis needs to be offered for easier understanding and unstructured Big Data makes this a challenge as well.

This paper offers a theoretical and analytic process to consider ways of analyzing Big Data and visualizing the analysis. To do this, it is important to offer a theoretical basis to consider the elements of unstructured Big Data.