iTrace-Toolkit: A Pipeline for Analyzing Eye-Tracking Data of Software Engineering Studies

Authors:
Publisher website
Google Scholar
Abstract
iTrace is community eye-tracking infrastructure that enables conducting eye-tracking studies within an Integrated Development Environment (IDE). It consists of a set of tools for gathering eye-tracking data on large real software projects within an IDE during studies on source code. Once the raw eye-tracking data is collected, processing is necessary before it can be used for analysis. Rather than provide the raw data for researchers to analyze and write their own customize scripts, we introduce iTrace-Toolkit - a suite of tools that assists with combining different data files generated from iTrace and its IDE plugins (namely Visual Studio, Atom, and Eclipse). iTrace-Toolkit also provides the crucial mapping of the valid raw eye-tracking data to source code tokens and finally generates fixations (an important metric in eye-tracking for comprehension) using three commonly used algorithms based on distance and velocity of eye movements. iTrace-Toolkit keeps track of all participant data and tasks during a given study and produces a complete lightweight database of the raw, mapped, and fixation data that is standardized and ready to be used by statistical tools. A simple GUI interface is provided for quick access to filter the data after an eye-tracking study. iTrace-Toolkit also allows for the export of the data or subset of the data to text formats for further statistical processing. YouTube Video: https://www.youtube.com/watchv9j20sOANh8w