A Scoping Review of the Digital Transformation Literature Using Scientometric Analysis
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Number of Parts | 30 | |
| Author | 0000-0001-6013-7437 (ORCID) 0000-0002-5184-3812 (ORCID) 0000-0001-8907-330X (ORCID) 0000-0002-3867-7055 (ORCID) | |
| License | CC Attribution 4.0 International: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
| Identifiers | 10.5446/53707 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
| |
| Keywords |
11
00:00
Computer animation
01:16
Computer animation
Transcript: English(auto-generated)
00:00
Good morning, everyone. My name is Ebert van Veltroven and I will be presenting our paper titled a scoping review of the digital transformation literature using a scientific analysis. So let's start with introduction. What is digital transformation? Digital transformation is hard to define and there are many different definitions in the
00:24
literature. One way to look at it is as a continuously increasing interaction between digital technologies, business and society. This has a large impact on companies. Not only their products and services are changing, but also their business model, their business structure, the
00:44
market they operate in or even the leadership. Digital transformation has been a popular research topic since 2015. It has become quite broad and it also has quite an unclear terminology.
01:01
Like I said, the many different definitions, but also many different synonymous terminologies are used. This makes it hard to keep track of what has been done. So that is why we attempted a literature review. We use scientific metrics for this literature review.
01:23
Scientific metrics is about analyzing the bibliometric data of the literature. So this data is actually the metadata of scientific publications such as the keywords, the authors, the abstracts, the year of publication and so on. We can visualize this
01:44
bibliometric data set in graphs and this way we can see the structure of the research field. This method is objective because we only look at the data itself. We do no interpretation of the data. It is scalable and it is also easy to interpret.
02:05
So our methodology is as follows. We conducted a query on Scopus and on Web of Science to search for papers with digital transformation in the keywords or in the title. We search for papers after the year 1999 and we also search only for English conference and journal papers.
02:27
So in total we retrieved 1985 papers on Scopus and 1158 papers on Web of Science. We combined these two data sets and they had 716 duplicates. So in the end we
02:45
have a data set of 2427 articles with 10 features, so these 10 bibliometric data types. This data set was then pre-processed. So first we merged the two data sets together with a custom
03:01
Python script. Then we manually checked the titles of the papers and we removed all the irrelevant papers. Next we did some manual data cleaning such as fixing the missing values, fixing some spelling errors and merging synonyms such as some authors would say CIO and others
03:24
would say Chief Information Officer. We then built another Python script to move all the words to lowercase except for the abbreviations. We used a dictionary to remove all the stop words and another dictionary to move all the UK spelling to the United States spelling.
03:46
We then also merged all the acronyms and we did a few more manipulations. Then we did lemmatization based on the stem values. So what is this? Lemma is basically
04:01
the dictionary form of the word. For instance transforms, transforming and transformed are all they have the lemma transform while transformation has lemma transformation. So in the first run we created a dictionary with the lemmas of every word in the data set.
04:21
In the second run we did the same but this time we created a dictionary for the stem of each word. So the stem is the short form it's from connects connecting connected the stem is connect and from connection is also connect. In the third iteration we then transform the stem
04:41
into the most popular lemma of that stem. So for instance imagine the data set has 60 instances of strategy and 45 instances of strategic. The stem of both of these words is strategy so we created or we changed all the stems into strategy the most popular lemma of
05:05
that stem. So why do we do this? This method has several advantages. Lemmas are actual words while stems are sometimes not actual words. This method also fixes the different spellings
05:22
for the same word and it also reduces the total number of the different words. For example after the lemmatization based on the stem values the total number of different words in the title got reduced from 2545 words to 2293. Let's take a look at the results then. So this
05:45
data set we just pre-processed are now being analyzed with a few figures. First let's take a look at the number of papers being published every single year. You can see a quick expansion from the year 2014 and onwards. The number of papers doubles annually. This shows that there
06:05
is an increased interest and importance in digital transformation research and one potential reason why it's so fast expanding is that similar terminologies are slowly consolidating into digital transformation such as before 2014 artists would maybe use IT enabled business
06:26
transformation and then afterwards they were starting to use digital transformation. Taking a look at the top outlets and the top conferences by citations and by number of papers
06:41
we can see that there are many different outlets and most of these outlets have no specific category for digital transformation. These are mostly information system journals and conferences. You also see that conferences are quite important both in number of papers and number of citations.
07:05
The top outlets are as you can see on top left, Sustainable Switzerland, Technological Forecasting and Social Exchange and the Journal of Business Research. Taking a look at the keywords over time we can see that once again the fast expansion of the field.
07:28
You can also see that digitalization and industry 4.0 are the most popular related keywords to digital transformation. You can also see some technologies in there that are quite popular
07:41
such as big data and AI and some economy concepts such as digital economy and business model. Moving on to the actual syntax, it looks like this. This is a large graph and it might be quite hard to see on your screen but this is a title cover network. Here every single note represents
08:08
a word of the data set. The size of the note represents how many times this word has been in the title of a paper. Here you can see the structure of the research field basically.
08:23
You can see digital transformation is mostly connected with organizations, with industry and technology and innovation and transformation. You can also see some typical title structures. For example we have organization and technology in the yellowish color, we have supply chain papers,
08:48
we have literature reviews, we have some research about health care, about public services, about management and strategy, about education, case studies, production and manufacturing
09:00
and economy and sector. This gives the first idea of how big this research field is. We did the same graph for the abstracts which you can see here and once again you can identify three main research hubs. We have the manufacturing, production and supply chain,
09:24
we have the development and implementation of technology and we have a large research field about strategy and management. Finally we also did this for the keywords. So here you can see that the keyword digital transformation has a high overlap with the
09:45
keywords digitalization and industry 4.0. We also see many different technologies in the blockchain and cloud computing and here again you can see some research hubs such as papers
10:06
about leadership, papers about business model innovation, about e-government, strategy, organizational culture, business agility, cyber-physical systems, education, manufacturing, health care, e-learning, IoT and AI and enterprise architecture. So these
10:30
networks can be used by people interested in the research field to get a quick glance of what the research field looks like. It's quite easy to see what has been researched and what
10:44
has not been researched yet and also how they relate to each other. Then we did something similar for the core artist networks so here every single node is an author and they are connected if they have papers together. You can see many many different research hubs and
11:04
they are not that well connected with each other. So we found more than 100 different research teams being active on digital transformation research and we also found a low level of collaboration. As you can see all of these little dots they are not connected with each other.
11:24
The largest collaboration network we found is this one. It's about 92 researchers from the Kremz University of Austria, the RWTH of Aachen in the Netherlands, National Institute in Argentina and the Warwick University in the UK.
11:45
So this is quite low the level of collaboration especially compared to other fields of this size. To discuss, in the literature digital transformation is not only about technology
12:01
and business but it's also about changes in people, changes in society and societal values and this was not visible in the results. So this might be an area where further research is needed. We also found some evidence of the often mentioned ill-defined terminology. We found
12:22
that 26% of the papers use one or more synonymous terms such as digital transformation, digitalization and industry 4.0 and we also found that this is quite broad. So maybe this field is so large because many papers use digital transformation not correctly.
12:42
Finally, the core author networks showed a lack of collaboration. This was quite unexpected and it is also low compared to other fields. The limitations of this study are as follows. So our data set can contain biases and different queries could also result in different outcomes
13:04
but the query was quite broad and quite general so we think this is quite limited. Then we also, the interpretation of these figures are also using the proxy that they, the title, the keywords and abstracts correctly represent the content of the paper.
13:24
So this was it. So thank you for your attention.