Dashboard Design - but is it art sweetie?

Dashboards are a collection of visuals that convey a coherent message. Many would say that the role of building a dashboard is the role of a data scientist. I would say it’s the role of a creator, after all, it is art sweety!
But even the most abstract of artists will have some rules or principles to follow otherwise the message or meaning they are trying to convey is lost. I feel the same about dashboards, so I thought I’d share some of my own rules: 

1) Create for the Consumer

​As data scientists we can be tempted to flex our analytical muscle and produce something that is ….. well….. scientific. But those who are consuming the analytics are much less likely to be a data scientists, they wouldn’t need us if they were. Curb the enthusiasm and always keep the consumer in mind. Build a dashboard that talks to their understanding of business process and their expectations of metrics. Keep it simple and it will succeed.

2) Its just like reading a book

A story is a collection of chapters and paragraphs. If those paragraphs all talk about something different and the chapters don’t follow a certain order, then the story just isn’t a story. The same applies to dashboards.

Let's assume a dashboard has a number of pages, these pages are chapters of a book and needs to tell one part of the story. Within each page there are visual metrics in tiles, each tile is a paragraph, which needs to be consistent in theme. Without this structure a consumer of a dashboard will not be able to easily understand what it is trying to say and won't be able to get any insights from it - which is it's entire purpose.

3) No time for tables

Controversial I know but if you want tables open a spreadsheet. The tiles on dashboards need to be able to communicate the metric at first glance. If a consumer needs to ask “what is this chart saying” then the tile has failed.

Do not compensate for failure by including a table. Tables take time to read and digest and quite simply take up too much space. By all means enable drill to detail functionality or exporting of underlying data but keep the tile visual. It's just more pleasing to the eye.

4) Pick the right visual

This may sound like a no brainer but too often have I seen a chart misused, overcluttered, over complicated or just plain unbearable!

Visuals are there to fulfill an analytical task such as comparing values, showing the percentage composition of a total, tracking a trend over time or understanding the relationship between value set. The visuals that are made available to us lend themselves well to that task – so pick the right one, convey a simple message and for the love of analytics limit your use of pies! ​

5) Filters find the facts

It may seem I'm contradicting my first principle by not building a tile with a consumer requested specific metric. I'm a firm believer in building consumer capability and encouraging them to interact with the functionality of a dashboard to find the fact they are looking for.

​Filters are a great way of making sure a single dashboard can tell multiple variations of the same story – use them, make your consumers use them, and their capability will grow.

6) The rule of six

My brain just can't take much more than that, nor can the page. I do believe chapter can be told in 6 paragraphs, there are after all only so many metrics you can squeeze out of a theme. Any more than that and you risk repeating yourself and discouraging the use of filters.

​This is also a real estate issue, if you start cluttering your page with more than 6 tiles you're going to lose the eye of the consumer and aint nobody got room for that.

---

These rules or principles are of course flexible, there's always room for artistic license for the purpose of story telling. If you're on a journey of dashboard creation and want some help in making sure they work for your consumers, then please do reach out. I'm always happy to share opinions.
​
P.s did you notice how many rules I have?