Table of Contents Scripts to analyse a the network in a body of literature (linux) Step 1. Search in Scopus Step 2. Use the scripts to reformat the resulting csv file to generate an overview of authors, papers and keywords. Step 3. Harmonise the result in Google Refine Step 4. Visualise and explore literature network with Gephi.

Scripts to analyse a the network in a body of literature (linux)

Step 1. Search in Scopus

• Use the document search to query for a number of search terms in titles, abstracts and keywords, where the search terms are separated with AND.
• Select all documents found and use the export function. Select the complete format and export to a csv file.

Step 2. Use the scripts to reformat the resulting csv file to generate an overview of authors, papers and keywords.

• Use generateAuthorOverview.sh to give an overview of the authors. The resulting .edges file contains an overview of the links between authors.
• Use generateDocumentOverview.sh to give an overview of the documents. The resulting .edges file contains an overview of the citation links between documents.
• Use generateKeywordOverview.sh to give an overview of the keywords listed. The resulting .edges file contains an overview of the keywords that are mentioned together. The keywords-count.txt files gives an ordered count of the keywords.

Step 3. Harmonise the result in Google Refine

• The result may need to be improved to resolve unmatched duplicates with Google Refine.
• Especially authors and documents need to be checked, because the same author and documents have different identifiers (for example ‘Nelson, R.’ and ‘Nelson, R.R.’).
• Use the Cluster and edit function to find similar values and determine which should be duplicates. There are various clustering algorithms implemented.
• Also use this function to combine various editions of the same publication into one document identifier.
• Change everything to lower case. Remove malformatted references.
• Use the Facet by Blank function to deselect empty cells.
• Export the result as a tab-separated file.

Step 4. Visualise and explore literature network with Gephi.

• All three lists are imported in Gephi to study the network.
• Check by hand for duplicate nodes and use the Merge nodes function.
• Remove erroneous nodes (such as commas only, or ‘from china’).
• Use the data explorer function to calculate general statistics of the networks and get an overview of the mostly cited papers.
• The network may be visualized best by using the ForceAtlas layout, by making the size of the nodes based on the degree.
• Export the resulting networks.