Retrospective: Data Visualisation- Learning through critique

James Northrop · 2019/03/09 · 4 minute read

Introduction

I’m not sure if the March 7 Data Science Melbourne Meetup’s last-minute speaker change classes as serendipity but the 3 replacement speakers presented highly commendable presentations at the last minute.

It’s unfortunate that the planned presenter, Cole Nussbaumer Knaflic from the US and author of Storytelling with Data, was unable to present at the Meetup due to ill health. I’m told that Cole is an inspiring speaker and well worth seeing if possible.

This retropsective focuses on 2 of the 3 speakers: Simon Beaumont from JLL in the UK and Fiona Gordon from JLL in Sydney.

Speakers

Simon Beaumont: JLL - UK

Simon leads a time of data analysts with a focus on delivering high levels of user experience (UX) and design consistency. In his words, he aims to fascilitate a continuous learning environment and culture.

As an example, Simon aims to cap dashboards to no more than 4 objects, thereby allowing plenty of white space.

An interesting point was that he “trains” his audience on how to interpret the dashboards. For example, he aims for charts to have a coloured triangle , with the colour prompting the user to:
1. Click on the triangle to open a new page.
2. Hover over the triangle for more information.
3. Click on the triangle to click through to an external page (e.g. for an external URL).

Simon brought my attention to Andy Kriebel’s wonderful Visual Vocabulary, which I’m sure that I’ll reference again and again in the future.

JLL’s visual standards are framed around 5 key dimensions:
* The dashboard layout.
* Interactivity (i.e. the coloured triangular cues). * Colours (i.e. consistency and cues). Simon used the BBC News and BBC Sport sites as examples.
* Fonts
* Formatting
* Chart types, with the reference to Andy Kriebel’s Visual Vocabulary.

Simon presented the “Viz Challenge” framework, established at JLL to bring data analysts together and inspire professional development. The challenge provides analyst with opportunities to be free of work pressures, enabling:
* Learning.
* Being creative.
* Supporting each other.
* Getting inspired by a personal passion of interest.
* Having fun.

Simon presented a tree map showing different opportunities staff have in professional development.
JLL learning tree map.

Another interesting idea from Simon was to try making a legend redundant through intelligent chart headings. Personally, I can see how this could work, and improve the aesthetic, though only in some circumstances.

Agile Data Visualisation.

Simon presented how his team is applying Agile methods to their work:
* Conducting daily huddles for feedback.
* Regularly engaging their customer and seeking feedback.
* Maintaining a work backlog and prioritisation process.
* Running iterative development.

JLL’s Agile approach to data visualisation

JLL’s Agile approach to data visualisation

JLL’s data visualisation goes through 3 stages:
1. Visualisation for submission is uploaded more than 24 hours before presenting.
2. A Webex team review is scheduled every Friday and open to the whole team.
3. Agreed refinements are made and approved work is published in a fortnightly release.

It is important to understand the framework for feedback.
1. Does the work adhere to design standards? (Design)
2. Are appropriate charts chosen? (Analysis)
3. Does the work answer business questions? (Insights)

Conclusion

Simon’s presented an inspiring message of how JLL enables a culture of continuous learning. I particularly appreciated these key messages as well as insights into how to adopt Agile into a data analysis team.

Fiona Gordon - JLL Sydney

Fiona (Fi) Gordon has recently started working with JLL Sydney after a period of time working for CommSec. Fi has developed the Vizchic.com website.

Fi’s mission is to enable “data literacy through gamification”, with the mantra to “embed data literacy in the organisation before you’re left behind”.

Fi defines data literacy as “the ability to gain meaningful insights from data”.

Fi advocates a 5-stage progression to data “rock-star” status with a highly structured process. A key enabler is advocating that staff spend 4h per week on professional development.

JLL Rock star steps

JLL Rock star steps