Eye Tracking Visualization Tool

About the project

This tool was created as part of a research lab project of the WeST Institute. Our goal was to develop new means to display eye tracking data. We started gathering our own test data by conducting an eye tracking study via the EYEVIDO Software and then build the tool to visualize our results. In the end we managed to not only implement basic visualizations such as heatmap and scanpath, but also added clustering and attention flow methods to better use the data we gathered. With an additional 3D mode, the user can also experience the time as a new axis.

You can find & test the tool here: https://eyevis.west.uni-koblenz.de/

Structure

css - all our CSS files. If you need to add a file, remember to add it to the index.html header

csv - our eye tracking dataset and Pyhton parser

image - images that we use

js - JavaScript files that we created or modified

js/global_vars.js - here we define global variables to use later in the code and transfer between people

lib - libraries that are in use

index.html - our single page website. Don't put any major JavaScript here. Only link script and css files.

Further explanation

The tool is displayed using one .html file and multiple panes to create the feeling of browsing a multi paged website:
- The Import data pane offers the possibility to try the tool with our collected test data or upload specific .csv files of eye tracking studies and their corresponding image files. We offer a Python Script to transform EYEVIDO SQL dumps into our required format.
- The Tool pane is the visualization part and lets the user chose which website of the uploaded data should be displayed and which visulization method should be applied. Each method comes with customization settings. We also offer general filter options to filter the displayed user data by metadata such as gender, age etc.
- The Help pane explains all available visualization methods and the conducted eye tracking study.
Most of the computing happens in the visualization.js and filtering.js:
- Upon selecting a website/screenshot which should be displayed, the visualization.js script takes care of loading the screenshot, the gaze and mouse data and the table of displayed users, aswell as assigning each user a unique color.
- filtering.js takes care of not only managing the filter options, but also displaying the right visualization options and calling the selected visualization method. The corresponding JavaScript file of the selected method will then take over to calculate and display the visualization on the canvas.

Project Team

Christian Brozmann
Chuyi Sun
Daniel Vossen
Marius Nisslmüller
Nadja Jelicic
Niklas Ecker
Oleksandr Kovtunov
Orkut Karacalik
Stefan Hill
Sophia Kramer

Under guidance of:

Dr. Chandan Kumar
Raphael Menges

Used libraries

License

The work is licensed under the MIT license. For more information please check the LICENSE file.

Shanksum/eyevis-tool