Search results

1 – 1 of 1
Open Access
Article
Publication date: 20 March 2017

Patrick OBrien, Kenning Arlitsch, Jeff Mixter, Jonathan Wheeler and Leila Belle Sterman

The purpose of this paper is to present data that begin to detail the deficiencies of log file analytics reporting methods that are commonly built into institutional repository…

5555

Abstract

Purpose

The purpose of this paper is to present data that begin to detail the deficiencies of log file analytics reporting methods that are commonly built into institutional repository (IR) platforms. The authors propose a new method for collecting and reporting IR item download metrics. This paper introduces a web service prototype that captures activity that current analytics methods are likely to either miss or over-report.

Design/methodology/approach

Data were extracted from DSpace Solr logs of an IR and were cross-referenced with Google Analytics and Google Search Console data to directly compare Citable Content Downloads recorded by each method.

Findings

This study provides evidence that log file analytics data appear to grossly over-report due to traffic from robots that are difficult to identify and screen. The study also introduces a proof-of-concept prototype that makes the research method easily accessible to IR managers who seek accurate counts of Citable Content Downloads.

Research limitations/implications

The method described in this paper does not account for direct access to Citable Content Downloads that originate outside Google Search properties.

Originality/value

This paper proposes that IR managers adopt a new reporting framework that classifies IR page views and download activity into three categories that communicate metrics about user activity related to the research process. It also proposes that IR managers rely on a hybrid of existing Google Services to improve reporting of Citable Content Downloads and offers a prototype web service where IR managers can test results for their repositories.

Details

Library Hi Tech, vol. 35 no. 1
Type: Research Article
ISSN: 0737-8831

Keywords

1 – 1 of 1