Primary supervisor

Yidan Zhang

Co-supervisors

This is a Summer Student Research Internship advert, running for 10-12 weeks over the summer break.

There is scope for this work to continue as an honours or minor thesis project afterwards. If you are interested in that, particularly the immersive visualisation side, please get in touch.

(Note that summer and winter student internships must be applied for here: https://www.monash.edu/study/fees-scholarships/scholarships/summer-winter)

--------------------------------------

Background
In our research we look at how people read and make sense of data when it is placed in the space around them rather than on a flat screen. Immersive environments let us present data at real scale and in real context, but they also raise questions that desktop studies cannot answer. Designs that work well on a monitor do not necessarily survive being viewed from an angle, from a distance, or while the viewer is moving through the scene, and we still have limited evidence about what works and why.

Answering that means running controlled studies with real participants in VR. Alongside the usual measures of accuracy, task time and what people tell us afterwards, we can also capture where participants look and how they move through the environment. We have Pupil Labs eye tracking available for use in VR (https://pupil-labs.com/products/vr-ar), which adds a useful layer here because it shows what people attended to and what they missed, not just whether they got the answer right.

Aim/outline

Over the summer we want to build and run a VR user study. The direction we are currently leaning towards builds on our earlier work using a space-time cube approach in VR. There we tested how effectively people read encodings such as colour and area when the data was mapped onto a single building and the viewer stayed in one place. The next step we have in mind is to bring movement into the picture, and possibly to spread the data across multiple buildings, so that people have to move through the environment and read the same information from changing distances and angles. Task performance would tell us whether the encodings still hold up, and gaze data would help explain why.

None of this is locked in yet, so there is room to shape the specifics with us and to bring in your own interests along the way.

Your role

  • Build the study application in Unity (If we go with the encoding study, there is an early prototype from earlier work to start from, so you would not be building from scratch, but expect to port and extend it);
  • Set up the Pupil Labs eye tracker as one of the measures, including calibration and a sanity check on accuracy;
  • Set up clean logging of task performance, gaze, head pose and position in the scene, properly time synchronised;
  • Prepare the study materials, such as participant instructions, training tasks and questionnaires;
  • Recruit participants and run the study sessions with support from the team;
  • Analyse the data and write up what you find.

Required knowledge

We are looking for a 3rd or final year undergraduate or Masters level student with solid VR development skills and an interest in how people read and use visual information.

Unity development with C# is essential. Beyond that, ideally you have:

  • Interest in VR, HCI and empirical research with people;
  • Experience designing or running user studies with human participants, particularly in VR;
  • Data analysis skills in Python or R, and at least basic knowledge of research methods and statistics;
  • Some experience with eye tracking hardware or software, ideally Pupil Labs, though we can teach this;
  • Some familiarity with data visualisation and visual perception, for example how effectively people read different encodings;
  • Familiarity with GitHub and collaborative coding workflows;
  • Ability to work independently as well as within a team.

A couple of practical points. The work is based on campus so you can use the lab equipment, you should be comfortable spending time in a headset, and you will be guiding participants through calibration and study sessions, so being at ease talking people through things matters.

What you get out of it

A VR study you built and ran, a dataset you collected yourself covering task performance and gaze behaviour, and a written report. Depending on how the study lands, there is potential for this to develop towards a publication.