Reconfiguration Strategies with Composite Data Physicalizations


  • Conference: CHI 2021
  • Type: Full paper
  • Goal: Investigate the relation between users' reconfiguration strategies and the physical structure of physicalizations.


Kim Sauvé, Lancaster University
David Verweij, (then) Newcastle University
Jason Alexander, University of Bath
Steven Houben, (then) Lancaster University






Composite data physicalizations allow for the physical reconfiguration of data points, creating new opportunities for interaction and engagement. However, there is a lack of understanding of people’s strategies and behaviors when directly manipulating physical data objects. In this paper, we systematically characterize different reconfiguration strategies using six exemplar physicalizations. We asked 20 participants to reorganize these exemplars with two levels of restriction: changing a single data object versus changing multiple data objects. Our findings show that there were two main reconfiguration strategies used: changes in proximity and changes in atomic orientation. We further characterize these using concrete examples of participant actions in relation to the structure of the physicalizations. We contribute an overview of reconfiguration strategies, which informs the design of future manually reconfigurable and dynamic composite physicalizations.


We designed 6 exemplar physicalizations, informed by the well-known physical bar charts often used in physicalization research. We ensured that these included various visual obstructions, gradual and abrupt height differences, and both clear and ambiguous clusters.


We presented each of the 6 physicalizations from four sides and asked 20 participants to reorganize the layouts according to the clusters they perceived. We introduced two levels of restriction to be able to generalize across systems of different interaction possibilities, in total this resulted in 480 study trials. In the restricted phase 1 we would ask participants how they would make the groups they identified more distinct by changing one object, and in phase 2 how they would do this without any restrictions.


In general, we observed two main reconfiguration strategies: changes in proximity and changes in atomic orientation, to create visual consistency and reduce ambiguities. Taking a closer look at the reconfiguration characteristics, we observed different approaches within the two main strategies, that could appear individually, or in combination.

Proximity changes

Within proximity changes we observed (i) cohesion changes, for example when moving objects within each cluster closer together, and (ii) separation changes, such as moving complete clusters further apart.

Cohesion change

Separation change

Atomic orientation changes

Within atomic orientation changes, we observed (i) rotations within the plane, (ii) rotations from plane to space, and vice versa, (iii) from space to plane.

Rotation within plane

Rotation plane to space

Rotation space to plane

Reconfiguration characteristics per physicalization

Looking at the reconfiguration characteristics per physicalization, our data shows that exemplar physicalizations with a constant internal structure are likely to involve either none or many changes (phys1 and phys2). Physicalizations representing overlapping clusters are likely to motivate a change of one or more centrally positioned objects (phys3 and phys5), and physicalizations with objects in mixed orientations likely result in rotation changes to integrate and/or differentiate their atomic orientations (phys4 and phys6).