Sci2 Manual : 5.2.6 Mapping the Field of RNAi Research (SDB Data)
This page last changed on Apr 01, 2011 by dapolley.
The data for this analysis comes from a search of the Scholarly Database (SDB) (http://sdb.cns.iu.edu/) for "RNAi" in "All Text" from Medline, NSF, NIH and USPTO. A copy of this data is available in 'yoursci2directory/sampledata/scientometrics/sdb/RNAi'. The default export format is .csv, which can be loaded directly into the Sci2 Tool.
To view the co-authorship network of Medline's RNAi records, go to 'File > Load' and open 'yoursci2directory/sampledata/scientometrics/sdb/RNAi/Medline_co-author_table(nwb_format).csv'_ in Standard csv format. SDB tables are already normalized, so simply run 'Data Preparation > Extract Co-Occurrence Network' using the default parameters: According to 'Analysis > Networks > Network Analysis Toolkit (NAT)', the output network has 21,578 nodes with 131 isolates, and 77,739 edges. Visualizing such a large network is memory-intensive, so extract only the largest connected component by running 'Analysis > Networks > Unweighted and Undirected > Weak Component Clustering' with the following parameters: Make sure the newly extracted network ("Weak Component Cluster of 6446 nodes") is selected in the data manager, and run 'Visualization > Networks > GUESS' followed by 'Layout > GEM'. A custom python script has been used to color and size the network in Figure 5.34.
To visualize the citation patterns of patents dealing with RNAi, load 'yoursci2directory/sampledata/scientometrics/sdb/RNAi/USPTO_citation_table(nwb_format).csv'_ in Standard csv format and run 'Data Preparation > Extract Bipartite Network' using the following parameters:
Then select "nodes based on ->" in the Object drop-down box, "bipartitetype" in the Property drop-down box, "==" in the Operator drop-down box, and "cited_patents" in the Value drop-down box. Press "Colour" and click on blue below. Repeat the previous steps, but change the Value to "citing_patent" and select the color red. Now press "Show Label". The resulting graph should look like Figure 5.35.
The SDB also outputs much more robust tables, for example 'yoursci2directory/sampledata/scientometrics/sdb/RNAi/Medline_master_table.csv'. This table includes full records of Medline papers, and will be used to find bursting terms from Medline abstracts dealing with RNAi. Load the file in Standard csv format and run 'Preprocessing > Topical > Lowercase, Tokenize, Stem, and Stopword Text' with the following parameters: Select the "with normalized abstract" table in the Data Manager and run 'Analysis > Topical > Burst Detection' with the following parameters: View the file "Burst detection analysis (date_cr_year, abstract): maximum burst level 1." There are more words than can easily be viewed with the horizontal bar graph, so sort the list by "Strength" and prune all but the strongest 10 words. Save the file as a new .csv and load it into the Sci2 Tool as a standard csv file. Save and view the resulting PostScript file using the workflow described in section 2.4 Saving Visualizations for Publication.
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
![]() |
Document generated by Confluence on May 31, 2011 15:16 |