Actions per Minute

This visualisation shows, for each participant, the number of actions performed per minute of specific actions. We do not include Details on Demand in this visualisation.

To see amount of brushing performed, see Private vs Shared Brushing Modes.

The outlier for the Grasped action was a specific participant, P23, who constantly made minor adjustments to 3D visualisations in front of him in order to rotate them. We filter out this participant for the purposes of readability, which we also do in the manuscript.

Based off our observations, there are valid reasons for some of the differences between the two study parts:

• Part A performed more Attached actions as many of them pinned 2D visualisations onto the walls. 3D visualisations in Part B were not conducive to being attached to walls, and little participants used the table.

• Part B Created more visualisations as participants had to tear out 3D visualisations in order to rotate them—a consequence of our panel design decision. Note however that the difference in visualisations Destroyed per minute is similar between both Parts, as participants equally found the "throw-away" metaphor for destroying visualisations to be difficult to perform.

• Part B saw more Grasping actions as participants grabbed and rotated 3D visualisations in order to view them from specific angles, rather than orbiting around them. This is in contrast to 2D visualisations in Part A, which effectively do not need to be moved and/or rotated as much.

• Part B saw very slightly more Transfer actions than Part A, as the larger size of 3D visualisations meant that they would accidentally "clip" onto panels more often, causing this transfer functionality to trigger.

Code

Based on StatisticsData.

We first perform further wrangling to normalise the data by time:

Then we plot using ggplot2 for the manuscript version:

And the version for the supplemental material (seen on this page):