Eye Movement Scanpaths for Autism Detection on Web Pages

Supplementary data for a paper presented at the 17th International Web for All Conference (W4A'20) proposes a novel autism detection approach. The dataset contains individual eye-movement paths used to evaluate a Scanpath Trend Analysis (STA) method. It also includes Python code to re-run the evaluation, supporting reproducibility.

Use Cases

• [Developing autism detection models] based on [eye-movement sequences recorded from web browsing].
• [Clustering and analyzing visual attention patterns] based on [scanpath trend analysis (STA)].
• [Reproducing and validating the proposed detection approach] based on [the included Python code and individual path data].
• [Studying atypical web interaction behaviors] based on [eye-tracking scanpaths from users with and without autism].

Strengths

• [Includes Python code for evaluation], supporting reproducibility and method verification.
• [Created to support a peer-reviewed conference paper (W4A'20)], indicating a research-grade purpose.
• [Based on the established Scanpath Trend Analysis (STA) methodology] referenced in multiple ACM publications.

Limitations

• [Key metadata is missing], including the number of rows, specific column names, and dataset size.
• [The temporal coverage and geographic origin of the collected eye-tracking data are not specified].
• [The dataset's scale and demographic diversity of participants are unknown], limiting generalizability assessments.

Provenance

Source

Sukru Eraslan, University of Manchester

Collection Method

Likely collected via eye-tracking experiments during web browsing tasks, as the data consists of individual eye-movement paths (scanpaths).