A Python library has been developed for analyzing eye movement data.

Contemporary research actively employs machine learning and artificial intelligence to analyze large volumes of eye movement data. However, despite significant advancements in this field, there are issues that hinder the effectiveness of such methods. One of these challenges is the limited flexibility of existing software solutions. They often provide a restricted set of parameter settings, making it difficult to adapt to the specific demands of research tasks. Additionally, the integration of these tools with other specialized programs remains a weak point.

The Python library EyeFeatures, developed at the Laboratory of Social and Cognitive Informatics at HSE in St. Petersburg, addresses these issues and offers a user-friendly set of tools for working with eye movement data. It includes modules for processing and analyzing data obtained through eye trackers — devices that record eye movements while performing various tasks.

Processing eye movement data is a complex process that involves several stages. Since the pupils move in a jerky manner rather than smoothly, sequentially focusing on specific points, the first stage of data processing involves identifying fixation areas. The second stage entails calculating metrics such as the average fixation duration and the average distance between points, which enables the creation of initial simple predictive or diagnostic models.

All stages of data processing can be performed using various modules of the EyeFeatures library. The flexible, modular approach allows for easy integration of eye movement data processing into existing research and commercial projects, from raw data to a ready predictive or explanatory model. For instance, applying the library in marketing research will enable the assessment of consumer reactions to advertisements. Eye movement analysis will reveal which specific elements capture the audience's attention the most.

Anton Surkov, project leader and junior researcher at the Laboratory of Social and Cognitive Informatics at HSE in St. Petersburg, states: “The library can be beneficial for researchers as it allows not only for the replication of what is already available in other software but also for the application of new algorithms and the creation of more powerful models for studies in areas such as marketing, cognitive process diagnostics, user interface development, and neurointerfaces (where control and interaction with the program occur through eye movement), combining components in ways that yield new results and enhance methodology.”

The development simplifies the data analysis process and accelerates the creation of predictive models, which is particularly useful in medical diagnostics, marketing, and the study of cognitive processes. The library has already been utilized in research for the strategic project “AI Technologies for Humans” and was presented at the international conference ECEM 2024 in Ireland.

The project is implemented within the framework of the strategic project “AI Technologies for Humans” (“Priority-2030”).