Nodelink vs Adjacency Matrice Study.

This repository contains supporting files for a study that compared Graph Visualizations (Node-link and Adjacency Matrice).

We evaluated interactive Node-link and Adjacency Matrice visualizations, using a between-subject methdology, with 557 distinct crowdsourced users. We used 1 real-world dataset, and evaluated 14 tasks.

Results from this study has been published in a conference paper that appeared in Graph Drawing (2017). And an extended version of this results has also been submitted to TVCG.

The sections below describe the files contained in each of the main directories.

Visualization Setups

This directory contains the visualizations (Node-link and Adjacency Matrice) and introduction files used for each of the 5 groups of the study.

Deployed versions of these study setups can be found at: http://ec2-34-229-95-54.compute-1.amazonaws.com/graphunit/nl-vs-mat-study/

Dataset

This directory contains the dataset used in the study. It contains the version of the Receipe dataset used for the study, as well as the positions file used to position nodes in the Node-link visualization.

It also contains the Airports dataset used for the second study.

Results

This directory contains the Quantitative results (Accuracy and Time) for each of the evaluated tasks and the Qualitative results (User Feedback, Window Sizes, etc.) for each of the 5 Groups of the studies.

Supplemental Material

This directory contains video demos of how the studies were presented to participants. It also contains spreadsheets of interactions we considered for this study, as well as a document of datasets we considered for this study.

Tasks

This directory contains all the tasks that were used in the study. The directory contains a ReadMe with a description for each task. Each task file is an XML file that contains the task instances that were used in the study. A given task instance (<question>), contains the nodes that were used in the task (<inputs>), and the correct answers (<answer>).

If a given <input> or <answer> involves more than one node, the nodes are separated with two semi-colons (;;).