Tobii Eye Tracking Analysis

This repository contains an end-to-end pipeline for analyzing Tobii eye-tracking data. The project is structured into modular pipelines, each focused on a specific type of analysis. Together, these pipelines provide tools for data cleanup, sequence analysis, clustering, visualization, and machine learning.

Repository Structure

tobii-eye-tracking-pipeline/
├── sequence-analysis/           # Sequence-based analysis (R)
│   ├── scripts/
│   ├── sample-data/
│   └── results/
├── cluster-analysis/            # Clustering gaze patterns (R)
│   ├── scripts/
│   ├── sample-data/
│   └── results/
├── visualizations/              # All visualization types (R)
│   ├── box-plots/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
│   ├── heat-maps/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
│   ├── coordinate-gaze-map/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
│   ├── aoi-gaze-map/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
├── machine-learning/            # Predictive modeling (Python)
│   ├── regression/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
│   ├── classification/
│   │   ├── scripts/
│   │   ├── sample-data/
│   │   └── results/
└── README.md

Each analysis or visualization folder includes:

• scripts/: Core code for that stage of analysis
• sample-data/: Example inputs to demonstrate usage
• results/: Representative outputs (plots, metrics, exports)

Tech Stack

• R

  • Data cleanup and preprocessing
  • Sequence analysis
  • Cluster analysis
  • Data visualization (ggplot2, etc.)

• Python

  • Machine learning (scikit-learn, pandas, numpy, etc.)
  • Model training and evaluation

• Tobii Pro Studio / Tobii data exports

  • Source format for raw eye-tracking data

Getting Started

1. Clone this repository:

   git clone https://github.com/USERNAME/tobii-eye-tracking-analysis.git
   cd tobii-eye-tracking-analysis

2. Explore each pipeline folder (e.g. sequence-analysis/). Each contains:

   • R or Python scripts for analysis
   • Sample datasets to test the workflow
   • Example outputs for reference

3. Dependencies:

   • R ≥ 4.0 (packages: tidyverse, TraMineR, cluster, ggplot2)
   • Python ≥ 3.9 (packages: scikit-learn, pandas, numpy, matplotlib)

Research Context

This repository was developed to support the study "Improving Data Visualization Comprehension and Sensemaking: An empirical study", which is currently under review at an academic journal.