RefViz  

By Dale Farris, Secretary Golden Triangle PC Club October 2003 RefViz is the super new program from ISI Researchsoft, the folks long known for their always valuable reference and citation managing programs, including Reference Manager, EndNote, and ProCite. Now, with RefViz, ISI Researchsoft provides a new text analysis and visualization software application for the researcher's desktop. RefViz is an amazing program. Just imaging being able to visually explore literature references. RefViz analyzes large numbers of references by thematic content and then presents an at-a-glance overview of the main topics discussed in the reference set. When I first saw what RefViz can do, I was amazed, and also very impressed with the power hidden beneath the program's very user-friendly interface. RefViz allows researchers to cast a wider net when searching the literature, broaden their understanding of a field, and easily focus on areas of interest. Used with ISI Researchsoft's other powerful programs EndNote, Reference Manager, and ProCite, RefViz provides an additional level of analysis to existing reference collections and lets users find the most relevant citations for their papers. With RefViz, you can now truly separate the wheat from the chaff, by literally seeing graphical images that represent the references and thumbnail, screentips of information about the references. RefViz will likely also point out references you might otherwise easily miss, despite your level of extensive research. RefViz Benefits Visually explore reference literature for major themes and topics Evaluate results in an interactive, visual landscape, and access complete reference detail in one click Retain important references otherwise lost when narrowing a search or skimming a list Import downloaded references directly into RefViz, or use reference collections from EndNote, ProCite, and Reference Manager Send references from RefViz to EndNote, ProCite, and Reference Manager, and publish your results faster How Does It Work? RefViz processes each dataset with an unbiased review of all terms found in the title and abstract/notes fields of references. It essentially groups the literature the same way you would stack papers. Two visualizations are provided to review the literature: the Galaxy and the Matrix. The Galaxy View RefViz analyzes large numbers of references by thematic content and presents an at-a-glance overview of the main topics discussed in the reference set. The Galaxy view organizes references into groups according to how they are conceptually related. The Galaxy view shows the results of searching for a term throughout the dataset. The Matrix View RefViz highlights associations between terms by presenting a review of which subjects tend to be discussed together in the literature (terms vs. terms), or an overview of the major concepts discussed across groups (terms vs. groups). Red cells at the intersection of the columns and rows indicate the corresponding terms are often discussed together in the reference set. Blue cells indicate few references contain both terms. White cells indicate no significant association in either direction. Publish Your Results Faster Create overviews of literature search results from popular databases, such as ISI Web of Science, OCLC, and PubMed, or bibliographic management software, such as EndNote, ProCite, and Reference Manager. Used together with EndNote, ProCite, and Reference Manager, RefViz provides an additional level of analysis to existing reference collections and lets users find the most relevant citations for their papers. Why Perform Text Analysis? Searching literature databases such as Medline, Eric, or PsycINFO is often an iterative query process that must balance the size of the final download with the potential loss of important information. Usually, the first query submitted is fairly generic and the results are overwhelming in number and cover a wide range of associated topics. A second, more refined query usually reduces the number of hits, but, depending on the subject matter, the resulting abstract set can be either too small to be informative or still too large to read and comprehend. Ideally, the larger information set will maximize learning about a topic of interest. However, it is difficult to remember both the broad concepts and the significant details for a large set of references. The text analysis and visualization capabilities in RefViz facilitate this process. With RefViz, you can get an overview of all of your references, and then gain in-depth knowledge of the topics of interest. While gaining a deeper knowledge of these topics, you can find out who else is working in your area, and read what they have published. You can also find trends over time and new topics emerging in your field. If you have bibliographic data in different software packages, for example, if different people in your group use different software, and you would like to combine all the references to obtain an overview of the entire reference collection, you can import data from each database and create a combined visualization. How Does Text Analysis Work? The process RefViz uses to divide a set of references into topic groups can be compared to the approach you might use to organize the references that tend to accumulate on your desk over time. Most likely, you would begin by reading the abstracts for each paper and placing them in stacks according to subject. RefViz sorts electronic copies of your important papers using a similar process: reading the contents, finding word patterns and their associations, and dividing the set of papers into groups based on subject matter. In addition to sorting papers, it also arranges the resulting stacks according to subject matter. This level of organization is analogous to arranging the stacks of similar papers on your office floor according to overlapping themes among the stacks. Instead of making one large stack for all the papers discussing a wide-ranging topic, such as medicine, you might decide to break that stack up into smaller groups based on subtopics, such as diagnosis, treatment, research, and public health studies. Groups of documents discussing the diagnosis of a certain disease might be placed next to the groups about research and treatment of that same disease. Groups of papers on the diagnosis, treatment, and research for another disease might be placed in a region distinct but somewhat close to the other disease set. The papers on public health would be separated from the other sets of papers. In the analogy of organizing papers in your office, the final step might be to place labels on top of each stack as a reminder of what they are about. RefViz also labels each group with the three terms that help distinguish it from the other groups in the reference set. These labels provide an easy to understand overview of the main topics in your reference collection. This within-stack and across-stack organization of your references is what you will see in the RefViz Galaxy visualization, but with a powerful advantage: interactivity. After the initial survey of the main themes in your reference collection, RefViz provides a suite of tools that allow you to query the literature set as a whole and find associations among keywords, concepts, and document groups. How Does RefViz Sort References Into Groups? RefViz uses mathematical algorithms to divide a set of papers into concept-based groups. Starting with a vocabulary list determined for each set of references, RefViz uses a statistical model to find key concepts. Unlike other literature sorting applications, RefViz defines key themes and concepts based on the context of the entire reference set rather than using pre-defined rules. RefViz sorts references into groups in the following steps: 1. Identifies words that represent key concepts, primary and secondary, for a reference set. 2. Uses the key concepts to create a mathematical signature for each document. 3. Applies standard mathematical clustering algorithms to partition the document set into groups of similar papers. The result is a much richer comparative analysis than a simple categorization of documents based on word counts. Coupling this with powerful visualization methods provides rapid insight into large reference collections. NOTICE REGARDING LEGAL USE OF DOWNLOADED REFERENCE DATA RefViz gives you the capability to visualize and analyze references from online databases and other sources. Some producers of online reference databases expressly prohibit such use and storage of their data, others charge an extra fee for a license to use the data in this way. Before you analyze references from a database, be sure to check the copyright and fair use notices for the database, carefully. Note that different databases from the same provider may have varying restrictions. RefViz allows you to import data from the following content providers and text formats: • Web of Knowledge -Web of Science, Current Contents, BIOSIS, CAB Abstracts, INSPEC, ISI Proceedings, ISI Meeting Abstracts • PubMed - MEDLINE format, saved as a text file • OVID – Medline, ERIC, PsycINFO • OCLC – Medline, ERIC, PsycINFO • CSA – Medline, ERIC, PsycINFO • ISIResearchSoft's RIS text format Additional data sources and formats may become supported. Check www.refviz.com. Target Customers RefViz is tailor-made for all serious researchers, university professors, graduate and undergraduate students, and anyone involved in serious research requiring use of online databases to support their research results and their publishing. The program thus is a definite must-have for all college and university libraries, and multiple copies should be considered for multiple library workstations, to support the number of expected users, once the word gets out among the research community that this super program is available. Price $299.95 list price with CD and manual $149.95 Special Introductory Price Contact the company for pricing for multiple licenses and for student/academic pricing. System Configuration Requirements Windows NT 4, ME, 2000, XP About Thomson ISI Thomson ISI is a business of The Thomson Corporation. Headquartered in Philadelphia, Pennsylvania, USA with offices worldwide, ISI provides essential, high quality Web-based information to over seven million researchers, information specialists, and administrators in diverse fields. The Thomson Corporation www.thomson.com, with 2002 revenues of $7.8 billion, is a global leader in providing integrated information solutions to business and professional customers. Thomson provides value-added information, software applications and tools to more than 20 million users in the fields of law, tax, accounting, financial services, higher education, reference information, corporate training and assessment, scientific research and healthcare. The Corporation's common shares are listed on the Toronto and New York Stock Exchanges (TSX: TOC; NYSE: TOC). Contact Donna Kirking Product Training Coordinator ISI ResearchSoft 800 Jones Street Berkley, California 94710 510-559-8592, Ext. 130 800-554-3049 FAX 510-559-8683 donna.kirking@isiresearchsoft.com www.isiresearchsoft.com