102 | JADE
ED DE QUINCEY ET AL
exam papers, screencasts and podcasts of some lectures, book lists, common teaching material, final year project documentation and relevant forms such as those for requesting extenuating circumstances, applying for ethical approval and for making general enquiries. Very few paper-based handouts are given to students so learning materials are only accessible to them via the intranet.
All student interaction with the CMS intranet is recorded in the form of server logs. When a user requests a file from a web server, an entry is recorded in a log file i. e. by loading a web page via a web browser, a user is making a request for a HTML file along with other files that are embedded components of that page such as images and videos; each of these file requests make an entry in a log. These server log entries contain information such as the name of the file that was requested, the address of the page that referred the user to the requested page, the IP address of the device that requested the file( this can indicate the location of the user), the time the file was requested and the username of the person requesting the file. As part of a previous study( de Quincey and Stoneham, 2015), functionality has been developed that takes this server log information and inserts it into a database, facilitating easier querying and analysis.
Server log data generated by 2,634 students across the School has been collected during the 2012-13 Academic Year with 2,544,374 interactions being recorded. Previous analysis( de Quincey and Stoneham, 2015) has suggested significant correlations between pairs of attributes on a Level 4 module called“ COMP1314: Digital Media, Computing and Programming”. COMP1314 is a 30 credit introductory module to computers and programming, delivered via weekly 2 hour lectures and 1 hour practical tutorial sessions across both semesters by 2 different lecturers. It is assessed by 2 pieces of coursework and an exam.
During the running of this module in 2012-13, there were 14,467 interactions with resources and pages on the CMS intranet related to the module by the 53 students who were still enrolled by the end of the module. Significantly high correlation was found between a student’ s final module mark and overall attendance at tutorial and lab sessions( r = 0.64( 1)) and a similar correlation between the final mark and their interactions with COMP1314 resources and pages on the CMS intranet( r = 0.63).
Here, a more holistic approach has been used, with the simple K-means clustering algorithm( MacQueen, 1967) being applied to a subset of the data generated by the programming component of the module delivered in the second semester( over 10 weeks),