A lament

Ryan Dunn
7 min readOct 29, 2017

There are undoubted challenges of communicating data. Most importantly the needs — including data literacy — of users. All too often data visualisation is considered as a supplement to data, an afterthought — all too often in the form of a thing people call a dashboard. This can lead to over simplification, dilution and — in the worst case scenario — confusion or loss of trust.

More empathy is required from data leaders to develop products which support users in clarifying the complexity that lies in the problems they need to understand. More sophisticated design is needed in developing data products which build confidence and literacy in data and visualisation.

We can do better for users.

Who the hell are you?

I lead a data science hub for the Department for Work and Pensions — the largest government department in the United Kingdom — responsible for providing security for millions of people at times of major vulnerability and turbulence in their lives. This is big and important.

My role is innovation. I co-ordinate a small research and development team. Our goal is, by combining data and digital, to contribute to improving outcomes and experiences for the people we support and — by making our organisation more effective and efficient — the people we work with in the department.

One of my aims is to raise expectations around the exchange of information — to make this conversational rather than transactional.

To achieve this aim, in the hub we work with users to design and develop digital data products for specific needs. Products that communicate data. Products which leverage the power of data visualisation.

Why are you writing this?

People sometimes call our products dashboards — or even fancy dashboards. This bothers me. Probably too much.

A month or so ago I ran a quick Twitter poll asking about a title for a blogpost I wasn’t sure I’d actually write. It got a bit of interest, a few comments and I had some good conversations with people at work about the subject.

I was very conscious that people I know, like and admire are working on things they — and others — call dashboards.

I didn’t — and don’t — want to undermine or disrespect their work in any way. I was concerned I couldn’t articulate the nuances of my thoughts about dashboards sufficiently well so I thought it was best to leave it.

However people kept calling our products dashboards — and it kept bothering me — so I wrote a few thoughts here in my #weeknotes and spoke a little about the subject at the Alan Turing Institute during a talk about our digital data products. Again these got a bit of interest, a few comments and conversations. Again I stopped short of writing an actual blogpost.

But I’ve been prompted enough now to suggest that it was worth putting my thoughts out there.

What do you know anyway?

Data visualisation has been part of my job for a few years now. During that time I’ve put the hours in and learned a lot. I care about the process and I value context.

Well executed data visualisation is an incredibly powerful tool for helping others to understand data — to help them ask better questions and make better decisions. I value the subtleties and complexities of the craft.

I continue to learn but remain in awe of the real experts. Many of whom I’ve been lucky enough to meet, work with, learn from and chat over a pint with.

I have talked on the subject myself and designed training — committed my time to trying to up the game — to improve data visualisation literacy — where I work and wider.

So I feel qualified enough to offer an opinion but I also know my limitations.

What’s your problem with dashboards?

In a nutshell — and bearing in mind the context of my work in digital government — I feel the word dashboard has become a lazy synonym for data visualisation. The concept of dashboard has moved on. The user need for information has evolved.

The word

What’s in a word. The board at the front of a 19th century carriage to stop mud or water which may splash — or dash — up from a horse’s hooves. The dashboard.

Windscreens were later placed above them and they became a handy place for gauges and minor controls. Vehicles and controls evolved and dashboards became home for an at a glance panel of indicators to monitor.

Computers came along. The term was borrowed for displays of the most important — key — information to monitor for a particular purpose. A single at a glance screen of simple charts providing a summary. The data dashboard.

So far so fine. Evolution of language, borrowing a term with the same conceptual meaning.

But computers evolved, data got bigger. This meant more metrics and breakdowns. This brought hierarchy, filtering and ‘drilling down’. Multiple screens now but still simple charts. The interactive data dashboard.

This is the important part for me. This subtle — but significant, and I think positive— shift from quite passive at a glance monitoring to much more active finding stuff out. Quite different things with the same name.

However a move from being told a story to being allowed to form your own — from explanatory to exploratory — presents a fundamental change in the expectations on our users —did anybody tell them?

This might seem minor. It probably is to the dashboard creators who feel they are offering the users the power to become more proactive than reactive. But this additional responsibility on users — the consumers — to have a conversation with the data, isn’t minor— especially in terms of data literacy and confidence.

Somewhere along the line the word dashboard has become adopted or associated with almost any visual display of numbers or interface to data. People see screens of visuals and think, dashboard. People hear dashboard and — I suspect —associate an expectation of at a glance monitoring. From a user perspective having one word with multiple potential definitions is confusing — what is expected of them?

Speaking to and working with users is so important. Defining the goal together, understanding the need, building the right thing.

The concept

There is definitely a place for more sophistication. Things have moved on, bigger data, more complex and interlinked systems and problems, more intelligent analytics and sophisticated visual design. In many situations, dashboard feels both too passive and over-simplistic.

There is also of course a place for simplicity. Summary visuals of summary data. Collect performance data, identify — key — indicators, record on a dashboard, monitor and take action accordingly. But the starting thought for a dashboard— using the at a glance monitoring definition of dashboard — is most definitely simplicity. This can obscure data rather than allowing users to clarify the complexity — to understand and unpick the intricacies.

Why limit the complexity that users need to understand? Why dilute communication? Instead, help users become more sophisticated consumers of information through digital data products with more complex (not complicated) visuals and embedded analytics.

Visual and interaction design are very important here. Considering the user experience of taking in and exploring information.

The user need

This is the most aspect- again bear in mind the context of my work in digital government. Dashboard has become a default —in how many instances is there a need as opposed to a want or mandate for dashboards — to monitor an at a glance panel of indicators.

The need — I hypothesise — is for better and deeper access to information that preserves the complexity of the problem and presents it in a way that it can be understood — and that is enjoyable and intuitive to interact with. The only way to find this out of course is to work more closely with users to understand their information needs.

This comes back to empathetic data leadership. Engaging users — who may not be confident with numbers — about data. Having the ability to translate user needs for data into something — perhaps something other than a report or a dashboard – that may be harder to develop but encourages a deeper understanding, that facilitates better questions and decisions.

Data literacy is about conversations, communication —it’s about meeting in the middle.

What do you suggest?

When there are real people and real lives behind the data, preserving context and complexity is incredibly important for decision making.

There is an ever growing space for products which use more sophisticated visual and interaction design to present information in ways which allow users to be proactive rather than reactive, or even passive, when it comes to data.

More data and better technology and design offer opportunities to develop products which can really narrow data literacy gaps when done well. They can of course have the opposite — very damaging — effect when done poorly.

So before defaulting to dashboard — have an open and iterative dialogue with your users, understand how they feel about data, talk with them about what they are trying to understand then work together to design an engaging, useful and usable solution.



Ryan Dunn

Data Science Hub Lead @DWPDigital. These are my personal thoughts.