Automated usability analysis and visualisation of eye tracking data

Abstract

Usability is a critical aspect of the success of any application. It can be the deciding factor for which an application is chosen and can have a dramatic effect on the productivity of users. Eye tracking has been successfully utilised as a usability evaluation tool, because of the strong link between where a person is looking and their cognitive activity. Currently, eye tracking usability evaluation is a time–intensive process, requiring extensive human expert analysis. It is therefore only feasible for small–scale usability testing. This study developed a method to reduce the time expert analysts spend interpreting eye tracking results, by automating part of the analysis process. This was accomplished by comparing the visual strategy of a benchmark user against the visual strategies of the remaining participants. A comparative study demonstrates how the resulting metrics highlight the same tasks with usability issues, as identified by an expert analyst. The method also produces visualisations to assist the expert in identifying problem areas on the user interface. Eye trackers are now available for various mobile devices, providing the opportunity to perform large–scale, remote eye tracking usability studies. The proposed approach makes it feasible to analyse these extensive eye tracking datasets and improve the usability of an application.