Theorizing Supply Chains with Qualitative Big Data and Topic Modeling

• DOI: http://doi.org/10.1111/jscm.12224
• Author: Nahyun Kim,Pratima (Tima) Bansal,Jury Gualandris
• Date: 01 April 2020
• Published date: 01 April 2020

THEORIZING SUPPLY CHAINS WITH QUALITATIVE BIG

DATA AND TOPIC MODELING

PRATIMA (TIMA) BANSAL, JURY GUALANDRIS , AND NAHYUN KIM

Western University

The availability of Big Data has opened up opportunities to study supply

chains. Whereas most scholars look to quantitative Big Data to build theo-

retical insights, in this paper we illustrate the value of qualitative Big Data.

We begin by describing the nature and properties of qualitative Big Data.

Then, we explain how one specific method, topic modeling, is particularly

useful in theorizing supply chains. Topic modeling identifies co-occurring

words in qualitative Big Data, which can reveal new constructs that are

difficult to see in such volume of data. Analyzing the relationships among

constructs or their descriptive content can help to understand and explain

how supply chains emerge, function, and adapt over time. As topic model-

ing has not yet been used to theorize supply chains, we illustrate the use

of this method and its relevance for future research by unpacking two

papers published in organizational theory journals.

Keywords: qualitative research; Big Data; topic modeling; complex adaptive systems;

networks

INTRODUCTION

The advent of digital technologies has opened up

opportunities for researchers to gather large volumes

of data, often called Big Data,

1

to theorize supply

chains. Quantitative Big Data can inform supply chain

theory because the large volume of granular data can

describe all aspects of supply chains, including cus-

tomer Web site visits, customer sentiments from social

media, varying service levels, or evolving contractual

ties and material flows in a supply chain (Miller, Gan-

ster & Griffis, 2018; Mishra et al., 2018; Park, Bellamy

& Basole, 2018).

Supply chain researchers thus far have largely

focused on quantitative Big Data, with qualitative Big

Data being largely overlooked. However, qualitative

Big Data can reveal how supply chains emerge, func-

tion, and adapt, complementing insights garnered

from quantitative Big Data. Qualitative Big Data offers

rich, contextualized information about multiple,

diverse organizations and their complex connections.

Whereas quantitative data often require researchers to

predetermine constructs and operationalizations, qual-

itative data permit researchers to induce insights,

which can reveal new constructs and potentially new

relationships among constructs. By iterating abduc-

tively between qualitative Big Data and existing the-

ory, researchers are given windows into theorizing

supply chains that would have otherwise been closed

with strictly quantitative data, such as the emergence,

functioning, and adaptation of unexpected interorga-

nizational structures and behaviors.

One of the strengths as well as challenges of qualita-

tive Big Data is that the data are incredibly rich of

contextual details, which makes them difficult to

reduce and analyze. Qualitative data include not only

text, but also videos, photographs, or sounds, which

can be scraped from Web sites, social media, news

articles, and open-source contracts. Machine learning

algorithms, especially those related to topic modeling,

can process these data to help researchers see patterns

that would otherwise have been impossible because

of the sheer volume of irreducible data.

In this paper, we first describe what we mean by

qualitative Big Data. Then, we describe topic model-

ing, which is particularly suited for extracting insights

from qualitative Big Data. This section is followed by

illustrations of the application of topic modeling in

two different studies in organizational theory journals,

with a description of how their analysis can be

All authors are contributed equally to the development of this

research paper.

1

We should note that although the word “data” is plural, we

treat the expression as a proper noun “Big Data,” which is capi-

talized and singular.

April 2020

7

Journal of Supply Chain Management

2020, 56(2), 7–18

©2020 Wiley Periodicals, Inc.