Converting Nataliya's CSV data from Scopus into networks I noticed that the records contain besides the author name also its Scopus Author Identifier (and for journals their ISSNs).
This provides a (at least partial) solution to the author's (and journal's) identification problem.
On the computer from the university/research domain (from outside using VPN), we visit the IZUM services page and click on the option Scopus. In Scopus, we enter the query determining the works of our interest. For example, the query k-truss AND ( graph OR network )
returns the following result
Selecting all hits (red square) and the option export we get
and afterward selecting the CSV option we get a window for selecting exported fields. We select all options in the first three columns, the Include references
option in the last column, and switch off the option Truncate to optimize for Excel
.
Clicking on the button Export we start the creation of the CSV file with our hits. In a similar way, we exported the hits also in RIS, BiBTeX, and plain text format. They are all available in the scopus-truss.zip file.
Up to 20000 hits can be saved at once.
Because Scopus provides besides the names of units (authors, journals, works) their IDs it simplifies the identification problem for them. We would need a program Scopus2Pajek
similar to WoS2Pajek
(GitHub/Bavla WoS2Pajek, Slides, Python) for converting Scopus data into Pajek network files - a combination of WoS2Pajek
and solutions in R.
There exists a program Scopus2WoS for the conversion of Scopus data in RIS format into WoS format. We selected the RIS format because it is very close to the WoS format. Unfortunately, it was a bad decision - in the RIS file, the IDs are not included.
In the conversion, we encountered the problem of replacing nonASCII characters with ASCII characters.