A new approach to web co‐link analysis

Abstract

Numerous web co-link studies have analyzed a wide variety of websites ranging from those in the academic and business arena to those dealing with politics and governments. Such studies uncover rich information about these organizations. In recent years, however, there has been a dearth of co-link analysis, mainly due to the lack of sources from which co-link data can be collected directly. Although several commercial services such as Alexa provide inlink data, none provide co-link data. We propose a new approach to web co-link analysis that can alleviate this problem so that researchers can continue to mine the valuable information contained in co-link data. The proposed approach has two components: (a) generating co-link data from inlink data using a computer program; (b) analyzing co-link data at the site level in addition to the page level that previous colink analyses have used. The site-level analysis has the potential of expanding co-link data sources. We tested this proposed approach by analyzing a group of websites focused on vaccination using Moz inlink data. We found that the approach is feasible, as we were able to generate co-link data from inlink data and analyze the co-link data with multidimensional scaling.

VINCENT: A visual analytics system for investigating the online vaccine debate

“…Websites that are plotted closer together share more online presence, while those plotted further away share fewer. Based on this map, polarity between the anti-and pro-vaccine websites was evident, similar to findings in previous related research [13]. All anti-vaccine websites ended up on the left side of the map, while all pro-vaccine websites are located on the right side with a space in the middle dividing the two.…”

“…The polarity of the vaccine debate is creating a clear divide and this has been revealed through both qualitative classification of inlinks [12] and quantitative co-link analysis [13]. The divide is having harmful effects on the health of the general population.…”

“…The list of 37 vaccine websites (Appendix 1) in VINCENT was created based on a list produced for a study on co-link analysis of vaccine websites which included a total of 62 websites [13].…”

“…Inlink data was collected from each website using MOZ's Link Explorer tool (https://www.moz.com/link-explorer). Diverging from some of the previous webometrics research using inlink data, which mostly investigated inlinks coming from pages (41,(43)(44)(45) and sites (13), VINCENT uses inlink data about the inlinking domains. Changes in September 2018 to the data provided by MOZ required us to adapt and examine the feasibility of using domainlevel inlink data.…”

“…Following similar data analysis techniques to that of previous MDS research [13], the inlink data collected on each website was used to create a similarity matrix, which is based on the number of co-links each website shared with one another. Using a computer program originally developed for a previous study [13], this co-link data was generated from the collected raw data. Using the output co-link matrix, the data was input into SPSS version 25 and an MDS analysis was conducted.…”