Preprocessing

Data Collection Naming Convention

The MultiplEYE preprocessing pipeline follows a standardized naming convention for data collections to ensure consistency across labs and languages. The pattern is:

[projectName]_[languageISOcode]_[countryISOcode]_[city]_[identifier]_[yearDataCollectionEnd]

For the data collection identifier MultiplEYE_SQ_CH_Zurich_1_2025, the breakdown includes MultiplEYE as the project name, SQ as the Albanian language code (see ISO 639 language codes), CH as the Switzerland country code (see ISO 3166 country codes), Zurich as the city name, 1 as the lab/experiment identifier, and 2025 as the year of data collection completion.

Eye-Tracking Preprocessing Pipeline

The main preprocessing pipeline handles the conversion and processing of eye-tracking data from the proprietary EyeLink format to analysis-ready data.

The pipeline processes data on a session level and consists of several key steps including EDF to ASC conversion, data parsing, gaze event detection, AOI mapping, and reading measures calculation. For detailed technical specifications of each step, including file formats and quality control procedures, please refer to the Technical Architecture section.

Input Data Structure

The expected eye-tracking data structure follows the MultiplEYE data folder organization. Each dataset contains eye-tracking sessions organized by participant and session identifiers, with stimulus information available in a separate folder containing AOI files. For the complete data structure specification, see the The MultiplEYE Data Structure section.

Running the Preprocessing

# Run full preprocessing pipeline
run_multipleye_preprocessing <data_collection_name>

# Run with uv
uv run run_multipleye_preprocessing <data_collection_name>

Output Files

The pipeline generates several types of output files including sample-level CSV files, gaze event files, AOI mapping files, and reading measures. Each step produces specific output files with standardized naming conventions. For detailed output file specifications and data quality reports, please refer to the Technical Architecture section.