I’m testing 2 formats of input file with the same data downloaded from Scopus: robust csv and ris scopus.
In order to transform original csv from Scopus into robust csv I’ve separated fields that are hosting multiple items with 3 successive stars.
So as the data has no difference, I hope to get the same maps created by 2 tested formats in Network Mapping script. Although there is some difference in auteurs’ homogeneous network depending on type of format (position of nodes and distance between them are not the same).
Besides, some columns from robust csv are not parsed into cortext db (“Art. No.”, “References”, “Year”).
Could you please check the absence of several columns and difference in maps?
Link for the project is below: https://managerv2.cortext.net/project/89690002702
I have run two Network Mapping analysis for Author-Author with raw as a proximity measure and no edge filter (Find the Optimal Proximity Threshold as No)
- And the networks are very similar. But, in the csv, there are some modifications made on the author names (e.g. no comma, space before some names…). These small differences make the two network and even more the edges proximity calculation a little different. If you want the exact same networks, you should not modify those strings.
- For the Year, during the robust csv parsing step, you should add the name of the time field (add “Year” for the “Time Field” parameter) in order to make CorText Manager able to build some temporal analysis. After the parsing the date will be nammed as ISIpubdate or date.
- Some of the fields in the csv are empty (e.g. Conference date…), so they do not appear in the CorText database.
- For “References”, just change the name of the fields. Open your CorText Manager projet, and look the scopus-60-doc-csv-lv. I have changed the name of References into: Cited_Ref.
I hope it helps!