Interactive Dataset

An Eye For AI Study: Dataset

A comprehensive study examining how developers interact with AI-powered code suggestions, analyzing attention patterns and decision-making processes through synchronized eye tracking and code suggestion data.

?

Research Team

Anonymized

Interactive Research Data

You are exploring the synchronized eye-tracking and code suggestion dataset of our "An Eye For AI" Study. The dataset contains 4444 code suggestions from 33 professional and student developers, and their fine-grained eye-tracking data. This interactive data visualization tool allows you to explore eye-tracking data collected during our research on human-AI interaction with code suggestions. You can filter and analyze data by different aspects in how developers interact with AI-powered code completions. The tool offers first offers a comprehensive summary of the dataset, and then allows you to explore the data in more detail. Finally, you can also explore single participants and their session data. The Key Metrics Explained section defines the terminology used for the tool and study.

Key Metrics Explained

Time Measurements
Duration:How long a suggestion stays visible on screen
Dwell Time:How long a participant actually looks at it
Suggestion Outcomes
Accepted:Inserted into code
Not Accepted:Not inserted into code
Actively Rejected:Rejected via ESC/click away

Loading research data...