Analysis of EZproxy server logs to visualise research activity in Curtin’s online library
| Date | 18 November 2019 |
| Pages | 845-865 |
| Published date | 18 November 2019 |
| DOI | https://doi.org/10.1108/LHT-04-2018-0050 |
| Author | Pauline Joseph,Aaron Justin Kent,Peter Damian Green,Matthew Robinson,Amanda Bellenger |
| Subject Matter | Library & information science |
Analysis of EZproxy server logs
to visualise research activity
in Curtin’s online library
Pauline Joseph
Discipline of Libraries, Archives, Records and Information Science,
Curtin University, Perth, Australia
Aaron Justin Kent
Curtin University, Perth, Australia, and
Peter Damian Green, Matthew Robinson and Amanda Bellenger
University Library, Curtin University, Perth, Australia
Abstract
Purpose –The purpose of this paper is to develop data visualisation proof of concept prototypes that will
enable the Curtin University Library team to explore its users’information-seeking behaviour and collection
use online by analysing the library’s EZproxy logs.
Design/methodology/approach –Curtin Library’s EZproxy log file data from 2013 to 2017 is used to
develop the data visualisation prototypes using Unity3D software.
Findings –Two visualisation prototypes from the EZproxy data set are developed. The first, “Global
Visualisation of Curtin Research Activity”, uses a geographical map of the world as a platform to show where
each research request comes from, the time each is made and the file size of the request. The second prototype,
“Database Usage Visualisation”, shows the use of the library’s various subscription databases by staff and
students daily, over a month in April 2017.
Research limitations/implications –The paperhas following limitations:working to a tight timelineof ten
weeks; time taken to cleanse noise data; and requirements for storingand hosting the voluminous data sets.
Practical implications –The prototypes provide visual evidence of the use of Curtin Library’s digital
resources at any time and from anywhere by its users, demonstrating the demand for the library’s online
service offerings. These prototype evidence-based data visualisations empower the library to communicate in
a compelling and interesting way how its services and subscriptions support Curtin University’s missions.
Originality/value –The paper provides innovative approaches to create immersive 3D data visualisation
prototypes to make sense of complex EZproxy data sets.
Keywords Academic libraries, User-centred design, Collection management,
Information resources management, Data visualization, EZproxy server logs
Paper type Research paper
Introduction
Curtin Library has a substantial data set of logged, authenticated use of its online library
collection, comprising databases, eJournals and eBooks dating from 2013. The EZproxy
software writesabout 30m lines a month, and this richdata set was accessible for thisproject.
Making sense of anddrawing meanings from the raw data,which is mainly in the textual
format of Uniform Resource Locators (URLs), is nearly impossible. The Curtin Library team
reportedan unsuccessful attempt to makesense of its EZproxy data sets10 years ago because
they foundit challenging to comprehendthe data using the then availabletools, as it all looked
similar, with no distinct observable trends in users’information-seeking behavio urs.
We acknowledge that the complex EZproxy data set provides a “snapshot in time that
captures something that moves and changes”quickly (Yau, 2013, p. 13). Hence, the difficulty
encountered by the library team only proved Yau’s (2013) observation that data are indeed Library Hi Tech
Vol. 37 No. 4, 2019
pp. 845-865
© Emerald PublishingLimited
0737-8831
DOI 10.1108/LHT-04-2018-0050
Received 16 April 2018
Revised 20 September 2018
18 January 2019
24 January 2019
Accepted 24 January 2019
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0737-8831.htm
The authors record thank you to Curtin University’s Office for Research and Development, for
sponsoring this internship and making this project possible.
845
Analysis of
EZproxy
server logs
an abstract concept to understand and they require context in the form of metadata (who,
what, when and how) to make sense. More importantly, we realised the requirement to use
visualisation to present data sets to our library audience, to make it easier for them to
explore and comprehend the deeper meanings it contains (Knaflic, 2015; Yau, 2013). This
view is supported by Tufte (2001), who describes the visual display of data as the “clear
portrayal of complexity”(p. 191) and our project aimed to apply modern visualisation
technology to this objective.
Given Spence’s (2014) description of visualisation as a metacognitive skill of “the
formation of a mental model of something”(p. 21), we sought an information visualisation
solution to comprehend the EZproxy data set. Chen (2017) states that “data visualisation
helps the human brain to process data faster and more effectively than text-based
information”(p. 24). Sviokla (2009) adds that visualisation enables efficient review and
processing of large data sets so that new knowledge can be derived and shared easily. Cruz
and Machado (2011) highlight the pivotal role visualisation plays in enabling synthesisation
of data so that insights gained from it are understood, transparent and convincing.
Scholarly findings about visualisations based on EZproxy data have been traditionally
static 2D or at the most 3D graphs, as in the work by Bhaskar et al. (2014), Coombs (2005),
Chan (2013), Grace and Bremner (2004), Lewellen et al. (2016), Murphy (2013, 2015),
Morton-Owens and Hanson (2012) and Sharman (2017). Dynamic visualisations are
uncommon in this field, but Archambault et al. (2015, p. 1) argue that a framework for
meaningful data visualisation has merit:
There are advantages to presenting data visually rather than as a set of flat statistics. Proper data
visualization facilitates the recognition of patterns and relationships to communicate a message in a
more compelling and interesting way. It allows the complexity of the data to be understood more easily.
We agree with Archambault et al. (2015) that the creation of a set of visualisations that is
dynamic and immersive would offer more insightful interactions with the data set, allowing one
to look at it in a way that normally is not possible. The creation of a 3D virtual space where one
could move around and view the data freely was envisioned to provide the library team with an
immersive encounter with the data set to explore their users’information-seeking behaviour.
The dynamic nature of the visualisation would allow user’s behaviour to be explored over time,
taking advantage of the detailed data contained in the data set that spans five years, and over
space using the geolocation information also contained in the data set. This visualisation of
researcher behaviour over time and space in an immersive environment would provide
unique insights into the research activity of users. In summary, we wanted to develop
information visualisation prototypes with interface functionalities that enhanced Curtin
librarians’interaction experience working dynamically with the different layers of information
embedded within the EZproxy data sets (Chen, 2017, p. 21).
An opportunity to work with the EZproxy data set was made possible with technical and
financial support from the Curtin Hub for Immersive Visualisation and eResearch (HIVE)
Internship Program, which allows a Curtin student to undertake a ten-week, full-time
investigation of the application of visualisation technologies to a discipline area. Interns
have regular access to the HIVE, are supported by its expert staff and supervised by a
library and information science discipline leader and the Curtin Library team; the latter were
the clients for this project.
In this paper, we describe an exploratory proof of concept project that aimed to visualise
the EZproxy data set to draw inferences about Curtin Library users’information-seeking
behaviour and collection use in the virtual/online environment. EZproxy offered a rich
source of data for analysis, containing a detailed log of Hypertext Transfer Protocol (HTTP)
requests processed through the library’s authentication servers. Another user identity data
set was merged with EZproxy to identify each users’profile: whether staff or student; their
846
LHT
37,4
To continue reading
Request your trial