Experimental analysis of the accessibility of drawings with few segments

This page serves as auxiliary material to our paper, entitled "Experimental analysis of the accessibility of drawings with few segments". Its main purpose is to make available the stimuli and collected data for the presented user study. The paper is available at arXiv.

Stimuli

The stimuli of our user study can be found on the four pages listed below. One page lists all the figures for the Aesthetic task for trees, one page lists all the figures for the Query task for trees together with the furthest pair and the path between them, one page lists all the figures for the Aesthetic task for sparse graphs, and one page lists all the figures for the Query task for sparse graphs together with the marked vertices and the shortest path between them.

Analysis

The analysis of the data was done with the programming language R. You may download our annotated analysis script below.

Data

The data we collected may be downloaded as five CSV files below. One contains some limited participant information; the others contain all responses for the four different tasks.

  1. Participants.

    Acolumns from familiar and on are given on a scale of 1 to 5, with 1 being the worst/most unfamiliar and 5 being best/most familiar.

    The background column is given as an integer, representing a 5-bit value. Their meaning per bit, from least to most significant bit:

    1. Math
    2. Computer science
    3. Human-computer interaction
    4. Visualization
    5. Other

  2. Tree Aesthetic Responses: original study - followup study - both.

    The layout algorithms are given as 'A', 'B' and 'C', with the following mapping to the algorithms.

    1. FewSegments
    2. Tidier
    3. Quad

    The balance of the trees are given as 'B', 'D' and 'W', with the following mapping to the tree balances.

    1. Balanced
    2. Deep
    3. Wide

  3. Tree Query Responses.

    The layout algorithms and tree balances are given the same way as in the Tree Aesthetic Responses.

    The 'furthestPairSource' and 'furthestPairTarget' columns contain the vertices corresponding to a furthest pair in the graph and the column 'furthestPairLength' contains their distance.

    The 'userPairSource' and 'userPairTarget' columns contain the vertices marked by the user and the column 'userPairLength' contains their distance.

  4. Sparse Graphs Aesthetic Responses.

    The layout algorithms are given as 'A' and 'B', with the following mapping to the algorithms.

    1. ForceDir
    2. FDFewSeg

    The type of the graphs are given as 'A', 'B', 'C', and 'D' with the following mapping to the graph types.

    1. ROME small
    2. ROME large
    3. Random small
    4. Random large

  5. Sparse Graphs Query Responses.

    The layout algorithms and graph types are given the same way as in the Sparse Graphs Aesthetic Responses.

    The 'shortestPathLength' column contains the length of the shortest path between the two marked vertices and the column 'shortestPathFull' contains all vertices on this path.

    The 'pathLength' column contains the length of the path selected by the user between the two marked vertices and the column 'pathFull' contains all vertices on this path. A '-' in the 'pathFull' states that the path is disconnected at this position.

    The 'clicklog' column contains the sequence of clicks on the vertices and edges and the Reset button together with the elapsed time since the start of the question.