a>

Phil Gough's Page

perceiving.info

User-Centered Design Process for Visualisation

April 8th, 2015

My PhD research has been comparing the different ways that a designer can communicate complex scientific information to non-scientists. This is an area of research that I am finding has a lot of activity, but little research. There are many research organisations, as well as individuals, that are using infographics, animated and narrated videos, and interactive art to communicate to the general public. This has been good for me, since there is a lot of room to do my research, and give names to a lot of things.

NEUVis and Design

I've been putting together an approach to NEUVis (Non-Exper User Visualisation) for my research. A modified approach will be useful in the HydratiCA d3 visualisaiton.

This image represents an approach to NEUVis using a user-centred design approach. I produced it at Sydney Uni in 2014 as part of my research to use in presentations.

This process will involve:

  1. Meeting the scientist which is using this visualisation tool for research, and discussing how he currently uses the visualisation, in order to find out what he needs from it, as well as what the data needs to show.
  2. Develop a description of the problems the design needs to overcome, based on these needs.
  3. Develop possible solution(s), and prototype.
  4. Test with the end-user and through heuristic evaluation
  5. Review the solution and redevelop (steps 2-4) if necessary
  6. Deploy the solution.


Heuristic and User Evaluation, User Needs

April 10th, 2015

Heuristic Evaluation

A heuristic evaluation is an evaluation of a system done against a set of defined principles. This is often done in conjunction with a user test. I use Nielson's Usability Heuristics.

#

Heuristic Name

Severity

Description

1

Visibility of System Status

High

User is not shown the timestamp being displayed. The input boxshows this, but it can be changed, and if the user presses enter, instead of clicking the button, it will not update. There is no direct feedback for this.

1


High

The Force-Directed Network graph is unnecessary. It can be better displayed.

Also shows blank data that is not connected to the rest of the 'network'.

1


Moderate

The word 'toggle' on the checkboxes is unnecessary.

2

Match Between the System and the Real World

Low

Chemical Equations are not shown the way that they normally would be—no subscript/superscript.

2


Low(?)

What are the reactions that are being selected in the 'data select' dropdown or the network graph. Why is it just 1, 2 etc.

3

User Control and Freedom

Low

No undo/redo

4

Consistency and Standards

Low

Remove the word 'toggle' from checkboxes

6

Recognition Rather than Recall

High

Equations are shown only as a number, rather than showing their formulae

8

Aesthetic and Minimalist Design

High

Info box should be hidden, and available on demand

"Settings" and "controls" should be separated, and only the controls visible, settings available on demand

9

Help Users Recognize, Diagnose and Recover From Errors

High

No error messages are displayed. There is no feedback when the system is working. Sometimes the display does not seem to update, and it is not clear what is happening (especially when changing the range in Histogram settings)

10

Help and Documentation

Low

There is no documentation

1

Visibility of System Status

Moderate

No values on axes in line graph or histogram. No connection between network graph and line graph/histogram. This could all be moved into one thing.

1, 9


High

There is no 'reset' in the 3D space—it would be good to go back to the default view without using MCP?

2, 4


Moderate

What to bar thicknesses relate to in the histogram? It doesn't make any sense to me.

User Needs

The user needs I have identified after the discussion with Jeff (et al.) on April 9th are:

How the visualisation stared - though I have moved the network to be a bit more useful.