INFRASTRUCTURE FOR LARGE SCALE HARVESTING OF WEBOMETRIC INDICATORS

Abstract

The main purpose of scientific rankings is not only comparison, but also stimulating the development of universities and research institutes, improving the effectiveness of their educational and research processes. However, existing webometric ratings are not suitable for this task because of the lack of the possibility of their analysis and verification. The solution is to change the approach to webometric measurements, to perform them based on the principles of regular collection and openness of source data. The article discusses the indicators used in the tasks of academic webometrics, their reliability and stability, substantiates the need for regular collection of the values of these indicators to improve the quality of data and analyze their dynamics. Based on the experience of implementing a project to research the Russian academic web space (http://www. webometrix.ru), problems are analyzed that arise when collecting the values of webometric indicators of sites of research organizations and institutions of higher education on a global scale with a frequency of at least 1 time per month.The author describes a distributed system for collecting values of webometric indicators and evaluates the necessary degree of parallelization of the process. The developed approach is universal for the tasks of collecting large amounts of information by the method of extracting data from website pages, and the collection of webometric data is also relevant for the tasks of analyzing other thematic segments of web space, such as library websites.