With the information explosion in the sciences showing no signs of abating, researchers must find new ways to cope with an ever-growing body of literature. RefViz is a text analysis and visualization program designed to sift through, analyze, and organize large sets of data. With the help of RefViz, researchers can identify potentially hot topics or areas in need of research in a given field.
RefViz can handle a maximum of 32,000 references, which can be imported from online databases such as PubMed, Ovid, and ISI Web of Knowledge, as well as from citation management programs. The data visualization tool can be configured in EndNote v.7 and higher and Reference Manager v.10 Network and higher. Reference Manager v.11 can be configured to export references directly into RefViz.
RefViz does not have a keyword or vocabulary file of its own. During the import process, RefViz basically “reads” through the titles and abstracts in a set of references looking for word patterns and their associations to identify primary and secondary keywords. The software then uses a mathematical algorithm to assign each reference a numerical value based on the keywords it contains. Finally, the software organizes similar references into groups based on the assigned values. To maintain consistency, references lacking abstracts are discarded automatically, so it is important to remember to include abstracts when downloading search results. This setting can be changed to accommodate reference sets that do not contain abstracts.
The RefViz window consists of four panes: galaxy/matrix, keywords/search, reference viewer, and advisor. The galaxy view (Figure 1) is basically a map of a set of references. Small dots represent individual references. Larger icons shaped like pieces of paper represent groups of similar references. The size of an icon illustrates the number of references belonging to a particular group: the bigger the icon, the greater the number of references in a particular group. Hovering the mouse pointer over an individual reference displays its title and the group number it belongs to. Hovering the mouse pointer over a group displays the group number, the number of references in the group, and the top three keywords in that group. Articles or groups focusing on similar topics are shown in close proximity to each other.
Figure 1.
RefVix galaxy view
The matrix view (Figure 2) displays a grid that shows the relationships between individual keywords and the groups that make up the set. Tabs at the top of the galaxy/matrix quadrant make it easy to toggle between the galaxy and matrix displays.
Figure 2.
RefViz matrix view
The keyword pane (Figure 3), located in the upper right-hand corner of the window, is divided into three columns: primary, secondary, and other descriptive terms. Each column lists keywords occurring in a set of references and indicates the number of references in which a particular word appears. Primary terms are those that RefViz has determined to be the main focus of a reference and are used to create group assignments during the import process. Secondary terms also influence group assignments, but not as strongly. It is very easy to tweak the keyword designations. For example, secondary keywords can be “promoted” to primary, and primary keywords can be “demoted” to secondary or “other” status. A thesaurus tool can be used to create synonyms. After changes have been made and the keywords are reprocessed, RefViz adjusts group assignments. These changes are reflected in both the galaxy and matrix views. The keyword pane also contains a search component that allows the user to focus in on a specific concept.
Figure 3.
RefViz keyword pane
RefViz's lower left pane contains the reference viewer (Figure 4), which lists references. The default display contains the group number, title, authors, and year. Additional fields can be added to the display, if necessary. The reference viewer can be sorted by clicking on a column header. Double clicking on a specific reference displays the complete reference in a separate window.
Figure 4.
RefViz reference viewer
The lower right pane contains the advisor, which provides context-sensitive help. If it is not needed, the advisor can be hidden, allowing the reference viewer to span the entire lower section of the RefViz screen.
A group of references with a set can be focused on by creating a subset view. This can be done by selecting references or groups in the galaxy/matrix views or by using the search tool in the keyword pane and then selecting “Create Subset View” from the file menu. Reference lists cannot be printed out directly from RefViz, but selected references can be saved as a text file or exported into citation management software for further use.
I ran a PubMed search on “evidence based medicine” and limited results to the years 2003–2004 and English language, creating a set of 4,120 references that I imported. RefViz sorted the references into 64 groups such as “teaching EBM,” “economic aspects of EBM,” “use of EBM in clinical settings,” and “nursing issues.” I used the thesaurus tool to create synonyms of the terms “EBM,” “evidence-based,” and “evidence-based-medicine.” Because the entire set of references dealt with the concept of evidence-based medicine, I demoted the keywords “EBM” and “medicine” from the secondary list to the other descriptive terms list, so they did not influence reference groupings. These adjustments changed the contents of a few of the groupings and rearranged the location of some groups in proximity to others in the galaxy view. Finally, I selected a group containing seventy-four references and exported them into an EndNote 7 Library.
RefViz is a valuable tool to help researchers get an overview of a broad field of study and to identify key concepts and ideas. It does not, however, replace the need for researchers to learn search skills. To locate articles on a specific, focused topic, it is better to do a detailed, refined literature search.