A free guide for learning
How to Plot a Term Map using VOSviewer
2023
Rizome Productions
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
Introduction
If you are embarking on a PhD (or any scientific research journey for that matter), you know that you will need to fairly quickly get to grips with the literature around your topic of interest and find a so-called ‘research gap’ to contribute to the production of new knowledge. You might know very broadly what you are interested in researching, but perhaps you are struggling to find exactly what part of this broad interest could use more and/or different lines of inquiry.
VOSviewer is a powerful tool that can be extremely useful for mapping the scientific landscape of your research interest and identifying research gaps. More than just simply mapping scientific literature using VOSviewer, the strategies applied for analysing these maps are key to generating an understanding of the structure of one’s chosen scientific landscape and thereby to finding research gaps. Simply describing what the map shows cannot lead one to arrive at research gaps, what is required is an analysis. It is only through analysis of the structure and contents of the map can you identify what is trending, what majority of the scholarship tends to focus on, how some of the minor and deviant voices in the literature propose investigating about your topic, and how you can draw novel connections between different streams/perspectives/lenses.
It is how you can plot a map using VOSviewer that forms the scope of this free guide. To learn about the different strategies available for plotting, manipulating, and analysing maps for visualising scientific literature and identifying research gaps, please refer to the full version of this guide.
So, what’s VOSviewer?
Before using a tool in the knowledge production process (or for any process), one must understand its capabilities (or lack of) as a means for setting expectations for how and in what ways it can be of help to you. So, let’s get into it.
According to the official website:
“VOSviewer is a software tool for constructing and visualising bibliometric networks.
These networks may for instance include journals, researchers, or individual publications, and they can be constructed based on citation, bibliographic coupling, co-citation, or co-authorship relations. VOSviewer also offers text mining functionality
that can be used to construct and visualise co-occurrence networks of important terms extracted from a body of scientific literature.”
In simple terms:
VOSviewer is used to visualise the so-called “structure of science”, by analysing bibliometric data. The structure of science refers to the relations between conceptual terms (words/phrases), authors and journals in a particular body of literature, which can be visualised using VOSViewer in the form of maps. This means that by plotting and analysing maps of bibliometric data you will be able to get a sense of the themes, terms, journals, and researchers for a particular scientific field, and how they interrelate to each other.
On a surface level, what can VOSviewer show you?
● terms, authors, institutions and journals that appear the most commonly in a body of literature
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
● how terms, authors, institutions and journals are linked to each other in a body of literature
● which terms, authors, institutions, and journals are highly cited in a body of literature
Through analysis of the above information, you will be able to:
● Quickly familiarise yourself with scientific landscapes
● Determine trends: historical, current, and emerging
● Identify research gaps
Let’s get started with learning how to plot your first VOSviewer map!
NOTE: If you decide to purchase the full guide (The Ultimate VOSviewer Guide: Visualise Scientific Literature and Identify Research Gaps), please make sure that you go through the first few chapters again as more detailed information and strategies regarding using VOSviewer to perform bibliometric analysis for identifying research gaps is provided.
Downloading VOSviewer
STEP 1: VOSviewer requires Java to run. Download and install the latest version of Java from the official website, www.java.com.
STEP 2: Once you’ve installed Java, go to the official VOSviewer website:
STEP 3: Navigate to the “Download” tab and choose the version of VOSviewer based on your operating system.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
Getting Data for VOSviewer
You will need to get data for plotting term maps on VOSviewer. Specifically, you will require metadata associated with journal articles, downloaded from a bibliometric database. This is because for plotting a term map, VOSviewer extracts terms from the title and abstracts of the publication metadata dataset you provide. In a VOSviewer map, a “term” can be a single word (e.g. “chicken”), short phrases (e.g. “kentucky fried chicken”) or an abbreviation (e.g. “kfc”).
More publications means more unique terms will be extracted.
Some bibliometric databases are free to use (such as PubMed, Dimensions and Lens.org) and some are not (such as Web of Science and Scopus). There is no “right” database; which database you use will depend on your field of research and the norms of your specific discipline. This guide recommends using Lens.org because it is free for personal use, is considered a broad database covering a lot of different research areas, has some cool in-built analytical tools and a generous export limit. Dimensions.ai is also free for personal use but has a lower export limit. If you have access to Scopus or Web of Science or are more comfortable searching in PubMed feel free to use these databases. At the time of writing, Google Scholar is not believed to be capable of exporting data for bibliometric mapping.
How many publications should I map?
When using VOSviewer, the number of publications you choose to include in your analysis is equivalent to the number of search results that are returned by your query in a database such as Lens.org, PubMed, Scopus etc. There is no definitive "right" number of publications to map:
it depends on your objective. For a detailed discussion on the number of publications you might wish to consider mapping, please refer to the full version of this guide.
Plotting your Map: Step by Step
To illustrate how to plot a map using VOSviewer, we will plot a map for the literature on ‘sleep’.
The search terms that you wish to use for your own research do not have to be as ‘broad’ as something like ‘sleep’. You can combine multiple keywords and use search phrases that align with your area of interest better. It is recommended that you actively follow the steps as you make your way through this guide and plot your own map.
Getting Data from Lens.org
STEP 1: Once you’ve navigated to Lens.org, On the right half of the landing page, click the
‘Scholarly Works’ tab in the ‘START YOUR SEARCH’ box. Then click ‘Structured Search’.
STEP 2: Click the dropdown menu titled “All Fields” and Select “Title, Abstract, Keyword or Field of Research”.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
There are two reasons why you should not search the full texts of articles:
1. Searching the titles and abstracts of articles makes your search results more focused.
The keyword you use to generate your search results could appear in the full text, even if the article is not primarily about the research area you want. Many articles that are not actually about sleep could contain the word “sleep” in the main body of the article.
Articles that contain the keyword in the title or abstract are much more likely to be relevant to the research area you want.
2. VOSviewer does not have the functionality to extract terms from the full text, only from the title and abstract.
STEP 3: Click the text field and type in your keywords. In this example, we will keep it broad and use “sleep”. Tick the box next to "Journal Article". This filters out book chapters, preprints and grey literature. To make the dataset more manageable, consider adding a date range to your query, in this example I used from 1 January 2019 to 1 January 2020. Click “search”.1
1 Please note that it is in this step that you should determine how many publications you wish to map as a means to control your dataset, given your objectives. There are various boundaries that you can draw around your topic of interest for making the dataset more manageable, and thereby further define the body of literature you wish to visualise. More information regarding this can be found in the full version of this VOSviewer Guide.
STEP 5: To export the file containing the metadata that VOSviewer needs to plot a map click
"Export".
STEP 6: Click the dropdown menu underneath the heading “Number of documents to include'' and select a number - make sure the number you choose is greater than or equal to the number of documents you want to export. Name your file and click "Export". Tip: name your file with a descriptive name so it's easy to remember what the file contains.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
Now we are ready with our dataset, we will move over to actually plotting a map in VOSviewer.
Plotting a Term Map
STEP 1: To open VOSviewer double click the “VOSviewer” application file located within the folder you downloaded from the official website. Once you have VOSviewer open, you will see the following interface. Click ‘CREATE’.
STEP 2: To plot a term map (as this is what’s relevant for visualising literature and identifying research gaps), select the third option: "Create a map based on text data”, and then click next.
STEP 3: Click "Read data from bibliographic database files". Click next.
STEP 4: Select the tab that corresponds to the database you have used to get your data.
Each database exports files in a slightly different format: make sure you select the right one. If the database you select in this step does not match the database you have used, VOSviewer may be unable to read your file. In this example, we used lens.org.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
STEP 5: Click the three dots and navigate to the location of the data in your computer.
STEP 6: My file was saved in my downloads folder, select it. Click OK.
STEP 7: Click "Next"
STEP 8: These settings are selected by default. Click next.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
VOSviewer will now extract the terms from the dataset. Depending on the size of the dataset, this can take up to several minutes.
STEP 9: Binary counting is selected by default. We do not have a thesaurus file yet, but you could build one in the next steps.
STEP 10: In this example, VOSviewer has extracted more than 200,000 unique terms from the dataset (shown below). It is important we determine the number of terms that we want in our map as it affects our analysis for understanding the broad structure of the field and consequently for identifying research gaps. You should generally aim for around 1000-1200 terms maximum in the final map. Maps with less terms are fine if your dataset is small.
We will control the number of terms in our map by adjusting the threshold for the minimum number of occurrences of a term. That is, the level of detail present in our final map depends on the term occurrence threshold that you set.
Important: In the next step, by default, VOSviewer selects the 60% most relevant terms that meet the occurrence threshold. Choose a threshold that includes around 2,000 terms (60% of 2,000 = 1,200). Once you are happy with the threshold, click next.
In this example, we will proceed 29 as the minimum number of occurrences of a term, 2004 terms meet the threshold.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
STEP 11: As a general rule of thumb, it is recommended you proceed with VOSviewer’s default settings for calculating relevance scores. That is, in this example, 1202 terms (60% of 2004) are selected for which relevance scores are calculated.
VOSviewer uses natural language processing to select “relevant terms”, which in this context means terms that have a specific meaning. Theoretically, the more general a term is, the lower its relevance score.
In our example dataset on sleep, some terms listed from high and low relevance scores calculated by VOSviewer are given in the table.
However, these relevance scores may be misleading: they are an indication of the “relevance” of the term to VOSviewer, not to you. It is a useful function to help you remove generic terms but what terms to include in your map or not should ultimately be up to you. For example, the term “cochrane library” (which is a medical database similar to Pubmed that contains medical literature) was assigned a relevance score of 2.25, while “sleep fragmentation” was assigned a score of 0.55. You could make a strong argument that “cochrane library” would add less value to your
map than “sleep fragmentation” if your goal was to understand the structure of research on sleep.
Once you click next, VOSviewer calculates the relevance scores for the number of terms to be selected.
You will see a window that is titled ‘Verify selected terms’, this is where you have a chance to clean your data for more refined maps. You can do so either by manually deselecting terms you do not wish to appear or by building a thesaurus. 2
Criteria for removing terms for cleaning your dataset:
What terms you remove is up to you but be careful: aim to remove general terms that out of context will not add value to your analysis. Removing terms that you are not interested in but are relevant to the field can skew the map and introduce bias. Removing terms is a skill that you will develop as you use VOSviewer more. What terms are irrelevant or not will also depend on your specific research area.
In general, you would probably remove:
● Terms relating to journals or databases e.g. j clin sleep med, scopus, cochrane library.
● Terms relating to the features of articles themselves e.g. abstract, title, conclusion, article.
● General terms relating to research design e.g. analysis, method, empirical, data.
● Terms that authors use to describe their results which are not helpful out of context e.g. mechanism, function, mediation, feature, activation.
2A VOSviewer thesaurus is a tool that, once built, streamlines your data cleaning and saves (a lot) of time for exploring a particular research field by iteratively plotting and analysing multiple maps: which is exactly what you should do for finding research gaps. If you do not make a thesaurus, you will need to manually deselect terms every time you plot a map.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
STEP 12: Once you’re happy with cleaning your dataset, click ‘Finish’ to plot your map!
STEP 13: Saving your maps
You should save your maps before closing VOSviewer. If you close VOSviewer without saving your map you will need to plot it again when you come back to it later. So, every time you plot a map, make sure you save it!
13.1: Under the “File” tab you’ll find options for Creating, Opening and Saving your maps.
Clicking the “Save” button will open a window as shown below.
It is recommended you save your maps as VOSviewer files; there are two types of VOSviewer files: map files and network files.
● Map files contain information about the items in the map: the labels, the sizes of the items and the clusters they belong to and so on.
● Network files contain information about the links that connect the items in the map: the link strengths, and what items are connected to each other.
When opening a VOSviewer map the map file is required and the network file is optional.
However, it is recommended you save both and open both as without the network file you will not be able to make as many adjustments/manipulations to the map and your analysis will be significantly harder.
13.2 Click the three dots as shown below:
13.3: Save the file in a location on your computer. Name the file (tip: add “map file” or “MF” or similar so you know which is which later):
13.4: Repeat the process for the network file and click OK.
San and Reece | RIZOME PRODUCTIONS | @PHDWITHSAN
STEP 14: To return to your maps and open a saved map, open VOSviewer and click on the
‘Open’ tab:
14.1: You will see this box open, which is identical to the box you saw when saving your maps.
Click the three dots as before
14.2: Locate the files you saved in your computer:
14.3: Open both the map file and the network file and click ‘OK’: