Visual Analytics around the Edges
Cloud created by:
8 April 2011
Presentation by Tony Hirst as part of the learning analytics sympsium at CAL 2011.
Once the preserve of system administrators, user activity statistics are no longer hidden away in server logs, but are increasingly easy to collect via third-party analytics services such as Google Analytics or by directly interrogating service APIs. Many commercial web operations use activity data to identify how their systems are being used, as well as apparent problems with them. Activity data collected over an extended period of time may also be used as part of recommendation systems that provide added value services to the users of the system.
We describe several informal attempts to gather and interpret user data in a variety of educational settings, from the formal to the informal, using Google Analytics and the Gephi open-source social-network analysis and visualisation tool.
In particular, we will review:
– the use of Google Analytics to analyse the usage of online course material in two ten-week-long Open University online courses over several presentations, treating the course materials as if they were ‘just’ web pages and reviewing the performance of the course materials as if they were ‘just another website’. We show how the reports identify progression of study through the course materials, as well as revealing important information about the length of time students appear to be spending in studying the online materials, an observation that can help inform the design of future course materials (Sheikh & Hirst, 2008). The analysis also demonstrates the care that must be taken in reporting website statistics, especially in the treatment of ‘average’ report statistics.
– the use of Gephi to analyse the structure of an online Twitter hashtag community that developed around the use of the PLENK2010 hashtag during the presentation of the PLENK2010 Massive Open Online Course, as well as the behaviour of a closed community of students engaged in an Open University design course. For PLENK2010, we will show how the network of friend relationships between regular users of the PLENK2010 hashtag on Twitter may be viewed as a set of connected clusters of users, and how each cluster represents a meaningful organisation of members. For the Open University design course, we analyse the behaviour of students who were commenting on and favouriting each others’ photos as part of an online design critique activity over an extended period of time, showing various ways in which a network representation can be used to capture different perspectives on the structure of the student community and its behaviour.
Sheikh, H., & Hirst, A. J. (2008). Library Analytics. Paper presented at Internet Librarian International, October 2008.