220.127.116.11 Document Co-Citation
Document Co-Citation Network
Scientometrics is a discipline which uses statistical and computational techniques in order to understand the structure and dynamics of science. Here we use ISI data from the journal "Scientometrics" and Science of Science and Innovation Policy (SciSIP) data from NSF Awards Search.
Download Scientometrics.isi. Load the file using 'File > Load' and locating the downloaded file. This domain level dataset is ideal for document co-citation analysis, as the scale is large enough that the resulting network will paint a fairly accurate picture of document similarity within the domain of scientometrics.
Select the dataset "2126 Unique ISI Records" in the Data Manager window and run 'Data Preparation > Extract Paper Citation Network'.
Two files will appear in the Data Manager window: the paper-citation network and the paper information table.
Select the "Extracted paper-citation network" and run 'Preprocessing > Networks > Extract Nodes Above or Below Value' with the following parameters:
The produced network contains only the original ISI records. Select the resulting file and run 'Data Preparation > Extract Document Co-Citation Network'. Then, select the network and run 'Analysis > Networks > Network Analysis Toolkit (NAT)'. There are 2056 nodes, 26070 edges and 775 isolates in the network. Run 'Preprocessing > Networks > Delete Isolates' to remove all the isolates. Because this network is too dense to lay out in GUESS, run 'Visualization > DrL (VxOrd)' with the parameters:
Next, select "Laid out with DrL" in the Data Manager and run 'Visualization > Network > GUESS'. Run the following commands in the GUESS "Interpreter":
Go to "Graph Modifier" and choose 'Object: nodes based on -> > Property: localcitationcount > Operators: >= > Value: 20 > Show Label'. See Figure 5.21.
Figure 5.30: Document co-citation network for Scientometric.isi in GUESS without DrL edge cutting (top) and with DrL (VxOrd) (bottom).
To see the log file from this workflow save the 18.104.22.168 Document Co-Citation log file.
22.214.171.124 Geographic Visualization
Using the dataset loaded from section 126.96.36.199, select '2126 Unique ISI Records' in the Data Manager. To find the latitude and longitudes of the locations of researches publishing in Scientometrics, run Analysis > Geospatial > Yahoo! Geocoder. Note that you will need a Yahoo Application ID in order to run this query. Enter your Yahoo Application ID (sign up for one here), "Address" for the Place Type, and the "Reprint Address" for the Place Column Name. Press 'OK'. This step may take several minutes to complete.
A new table will appear in the Data Manager labeled 'With Latitude & Longitude from 'Reprint Address'' containing all data from the initial table, plus columns with the Yahoo Geocoded Latitude and Longitude Coordinates. Select this file and run Visualization > Geospatial > Proportional Symbol Map using the parameters listed below.
Save and view the resulting visualization using the directions described at 2.4 Saving Visualizations for Publication. This will be the result:
Figure 5.31: Scientometrics.isi geospatial visualization. Circle sizes represent times cited, colors represent publication year.