Nasir Baba is a Faculty member with Usmanu Danfodiyo University and a fellow of the African Humanities Programme (AHP), ACLS.
Research work can seem isolated and lonely when there is only a limited supply of relevant related literature available to a researcher. Whereas access to the Internet may appear to have offered a solution to the problem of limited research literature, it may have opened another frustrating challenge. That is how to navigate the thousands, if not millions, of literature sources to sift out those that are most relevant to the research one is undertaking.
Imagine that you are writing a research article, you have only a few relevant sources to rely on and are at the risk of abandoning the research. What if you can use the sources that you already have to mine articles that lead give you more sources on the same content? Then "connected papers" helps with that. Just paste the title of the article and it will come up with all the papers related to that article.
A Visual Interface of Connected Papers Search Result
First, it was with Google Scholar
Google Scholar, for example, has two features associated with its search that enable researchers to track down literature sources that closely match the ones that they had already obtained and used. These features are the 'cited by' and 'related articles' functions that are below every source that appears in Google Scholar searches. Watch a Google Scholar tutorial here.
Much as these features of Google Scholar is helpful, I believe we all wished we had something much more systematic in the way it organizes the literature sources based on their relevance. This is where connected papers come in handy.
Each of the related papers will be represented by a node and the nodes are in different sizes. The more cited a paper is, the bigger the node representing it. Similarly, the darker the node, the more recent the paper is. Some papers are also nested closer than the others because more similar papers are closer to each other and have stronger connecting lines.
What "Connected Papers" does differently
With "Connected Papers" you only need to enter an article into the search box through several options that include its Digital Object Identifier (DOI) number, URL, or its title. Connected papers will build a graph giving you a network of similar papers on the same subject matter.
Let us assume you are writing on a subject on which you have only limited previous knowledge or resources. I suppose one of the first places you would begin your search from is Google Scholar. The search on Google Scholar may have produced a relevant article and you would want to get more like it. Copy the article’s DOI title or URL if it is on a repository like PubMed, arXiv, or Semantic Scholar.
Paste the DOI, title, URL into the connected papers search box and click the build a graph button. Notice how the percentage progress is displayed on the page as the site scurries the Internet for papers related to your search. When it is done it will display the list of related papers and a graphical network with each paper represented by a node. Watch a tutorial on connected papers here.
You will notice that the nodes are in different sizes because the more cited a paper is, the bigger the node representing it. Similarly, the darker the node, the more recent the paper is. Notice also how some papers are nested closer than the others. This is because more similar papers are closer to each other and have stronger connecting lines.
To the left of the page is a list of all related papers. As you hover over each title with your cursor, you will see its details displayed to the right of the page. The details displayed would include a summary of the article’s main contents.
A Visual Interface and List of Connected Papers
It is also possible to disaggregate the list of related papers into prior or derivative works. Prior works are papers that were most commonly cited by the papers in the graph. Derivative works are papers that cited many of the papers in the graph.
Prior and Derivative Works
You can expand the view of the list of related papers that are to the left of the page. When you do that, notice how some vital metrics are populated for each article including its authors, year of publication, and the number of times that it has been cited. Of particular relevance is the similarity index for each article as a measure of its relatedness to the original article. You can also download the entire list of related articles as a Bibtex file which you can then import into a citation manager such as EndNote, Mendeley, and Zotero.
It is also possible to disaggregate the list of related papers into prior or derivative works. Prior works are papers that were most commonly cited by the papers in the graph. This level of citation suggests that they are important seminal works on the subject that a researcher would need to get familiar with.
Derivative works are papers that cited many of the papers in the graph. This usually means that they are either surveys of the field or recent relevant works that were inspired by the papers in the graph. It is possible to download the list in both categories if you want to.
With the level of details on related papers and linkages established between them, connected papers provide a lay of the field to researchers interested in any subject matter. It informs the researcher about the most important works on a subject matter including those that may be easily missed because they were recently published. Overall, the site fills a critical gap for researchers as they work to situate their researches within the existing body of knowledge.
© 2021 Nasir Baba