Introduction
Earlier this year at work I gave a 30-minute flash Tech Talk during a Data Science Tech Tracks Day at work on ‘The What, Why, How and Who of Good Data Visualisation’. It was amazing seeing the night before that 117 folk at TR had signed up to listen, and maxing out at ninety during the event itself was a great result for a talk I have given a few times before in various guises during my time at Thomson Reuters.
In this article I want to touch upon a single element of that talk, that, although I saved it to the last few slides of my talk, I feel is perhaps one of the more important, if not the most important aspect of developing effective data visualisation and visual content. One that is enforced upon me, the Innovation Labs I worked in at the time, and the wider company of Thomson Reuters as a mandated demand, and has also been in my previous data visualisation roles, rather than a nicety to be considered at the end of a project if time is available and allows.
This particular aspect of the data visualisation process is very important to me, and nicely aligns with the title of this posting; a posting I have been wanting to write for almost a year now. Now I feel the time is ripe and right to do so, and also to reach out to those in those in the wider Medium community for comment and discussion.
The ‘Who’ of good data visualisation
For me the ‘Who’ element is far and away the most important of the talk I gave. Good, objective data visualisation requires input from a range on parties involved in the data’s journey, though this is largely dependent on, and informed by, the end goal of the prescribed visual content (for example, medium of output, target audience).
Ask yourself these questions; ‘Who does data visualisation?’; ‘How is data visualisation achieved and produced?’; ‘How and when is visual content deemed to be finished and complete?’; ‘Who inputted into your data visualisation/visual content?’.
Being a data visualisation specialist (so said my job title at the) a core requirement of my work was to produce final visual design products for projects and ad-hoc requests and deliverables such conference presentations and handouts/flyers, slide decks, publication images and infographics. This invariably steers my own contribution to projects towards their end phases. This is in addition to those contributions made by ancillary roles along the timeline of the project.
In all data visualisation roles I have filled, no data vis product I have or will produce in the future will be made by myself in isolation; no product should be created in isolation. Every product needs contribution from two or more contributory groups listed below; those involved will though largely depend on the structure, size and form of the company and project in question.
· Branding
· Journalism
· Publishing
· Proof readers
· Data scientists
· Data custodians
· UX designers
· Customers and clients
· Data visualisation experts/specialists
and … arguably … the most important …
· the target audience
I mentioned earlier that my work — and my role in Lab projects — is often found towards the end of a project timeline. In product production terms this is very true; but product production of visual content often also marks the convenient or enforced end of once phase or chapter of work and the start of the next at stages earlier in the project life-cycle. Because of the multiple sources of input I list above, I have always really liked to favour including as many of these as possible at these key milestone points, often in the form of review, sign-off or progress/update meetings, to reflect on the course taken to that point, and the direction the visual content should take in to the future. This has been an approach I chose to adopt early in my data vis career, has held true in all the job posts I have held, and so I continue applying this approach to this day, as it works for me. This should not be confused with opting to resort to a ‘scatter-gun’ approach with invites, inviting everyone and anyone in the hope of covering all bases and limiting impact of people ‘missing out’. But more so, that I appreciate and want to acknowledge that many different people and vital skillsets have contributed to the product up to that point in its lifecycle. Consequently, they will likely all have a genuine interest and need to seek affirmation as to where and how their data and input earlier in the build process of a visualisation has been used.
For me, data visualisation is a very emotive topic, and I realise it is a fundamental substrate layer on which the wider data science community resides and is often dependent to convey their findings and underlying stories within the data they use. People invariably like to see [their] data in more easily comprehensible, digestible formats than may be offered in just numeric tables and lists, and get will excited in having the opportunity to do so. I get great enjoyment and satisfaction from positive feedback and responses of customers who have requested visual work from me, as well as those who are involved in the earlier stages of data preparation and analysis, and just show a genuine interest and keenness to see their raw data in a different light. Checkpoint/check-in discussions allows this opportunity, and will also allow round-table discussion and sign-off of more monumental aspects of the visual content such as footnotes, titles, copyright legalise, as well as ensuring alignment with corporate branding and design standards. Doing so intentionally (I hope) cuts out a lot of miscommunication, misunderstanding and confusion.
But why have a put my role, that of Data visualisation experts and specialists, last in the bulleted list above? Not for a moment am I degrading these positions in designing the visual content for projects, but in some ways putting it near-last on the list goes some way to highlighting that I cannot do my job alone, and without contribution and assistance from a number of other roles within the typical corporate environment. No visual content can and should justifiably be created in isolation by data visualisation experts and specialists on their own.
Discussion
But why have I put the target audience last, and labelled them as the most important? Is the target audience not just one and the same as the product customer, and why are they [as] important as customers to the Labs in which I worked?
In my opinion, products should primarily be designed for the end audience, not the customer, business area or individual that requests them. Invariably visual content is aimed towards an intended end audience, be it the audience to a conference presentation, a printed publication’s readership, or perhaps a team manager reviewing a summary slide deck. They are one of many groups with a vested interest in any visual content produced, and are more often than not the group that need to absorb the information conveyed by the visual content, but simultaneously (and often unavoidably), that are not accessible to comment and critique during the production process. Often they are external to Thomson Reuters, external to the production process and not offered the means to critique and feedback to the production process to improve the product in question. Through adopting an ‘all inclusive’ approach to producing visual content, and attendance at review/check-point meetings, I feel this assures the greatest chance of achieving the most suitable visual product for the target audience in question.
Above all, I hope this article outlines that Data Visualisation is such a collaborative, yet subjective experience; it involves many groups, each with their own interests, influences and ultimate aims in the final product. However, arguably and perhaps most importantly of all, and at an implicit level unaware to themselves, is the target audience.
Data is important. But what matters is what you do with it
(Chief Economist, Hal Varian, Google)