Long-form Tableau Visualisations - The New Normal?

I've never made it a secret. I'm by and large against the principle of long-form visualisations.

My reasoning has been that while they're very useful to simply scroll through as an image on a mobile, there's a lot to consume when viewed in browser.

The hashtag #saynotolongform which I've sporadically used is designed to be provocative - There are some excellent examples of long & thin dashboards. But for me, this is sparking a real change in the way that the construction and consumption of this format - particularly on the side of a designer. And increasingly it's "data visualisation" which has taken the back seat with a more 'poster' style of dashboard taking the lead.

Edward Tufte talked about the data-ink ratio, but perhaps a new phenomenon with the rise of long form visualisations in Tableau is the data to image ratio. 

Lots of logos, images, shapes, custom lines, fancy borders/background images are married to stunning colours to create some simply gorgeous visualisations. But a lot of the information ends up being hidden in text; either with BIG STATEMENTS, large numbers or just integrated in descriptions.

The visual analysis which Tableau champions, where bodies of text and tables are replaced by the 'right visualisation for the message' - is this slowly disappearing in the public space, in favour of eye candy?

Money where my mouth is

So in lieu in finding out (if you excuse the phrase) how the other half live, for the Iron Viz Europe feeder contest I actively decided I wanted to create a long-form visualisation. And perhaps it showed why I wasn't a fan, as I really didn't enjoy the experience of building one. I've also sometimes found that some long-form visualisations are designed in such a way that when scrolling, I can forget or lose the underlying message - or worse, the story or the message that I thought a visualisation was telling is not the one the author is allegedly attempting to convey.

I feel as though, if Tableau are going to feature and focus upon long form visualisations in the public sphere, perhaps there should be an effort to also make it easier. This said, there's also scope I think for some developments on dashboarding in general, and updating this to fit the use cases people use Tableau for - both as part of their day-to-day jobs as well as for data hobbyism.

The interesting thing is that long-form seems to be something which has ramped up in the last 12-18 months - perhaps even coinciding with the 'Device Specific Dashboarding' feature which dropped along with Tableau 10.0 - And while I can see it's benefit for a mobile viz, I'd even go as far to say that I am yet to see a fantastic mobile dashboard, with rich amounts of analysis, insight & a clear story (And please, if you're reading this post, do send me your favourite long-form vizzes, and why you like them!)

But designing for the medium of consumption is something which is absolutely must be considered. This is where I feel that 'dashboard design' in itself, is not a skill. It's a part of a wider collection, and I think we as a community should stop using it as a blanket term for specific data visualisation decisions made with purpose. The current connotations just seem to be pure 'design' related, than about good dashboarding, especially when it comes to different approaches to storytelling.

Championing Long-Form

I think the difficult part of the Iron Viz contests is looking beyond the aesthetics of a visualisation (by this I mean using a lot of colour, text & images to assist/add context to the data) and focussing on analysis. This is where I think the data to image ratio comes into play - tables of data have now been replaced with text, large bold statements & other image-based clutter.

Indeed, design comes as one of the 4 sections that's being measured by the Tableau Public team - Design, Storytelling, Analysis & Overall Appeal being the categories - but it seems as though there's been a shift, or even a skew by the dashboard designers in the Community to use long-form to emphasise the design aspect, perhaps even more than the data visualisation.

Why is this?

In my mind, the reason for this is essentially because of the 'storytelling' section of Iron Viz. With over 60 entries for the first Iron Viz feeder, in order to stand out it can almost become a pre-requisite to create a visualisation which has a lot going on. By this I mean a dashboard which follows the 'flow' of a story, which each 'section' getting small nuggets of analysis.

Indeed, while browsing through the entries this morning, I noticed a few which had 2 charts on the dashboard. While I understand & am an active fan of 1-chart dashboards, my view is that the annotation layer (this post is great on this topic) should drive the storytelling.

What is the limit? When does it become a problem to have more images & headers over insight into datasets?

From 2 charts/sheets/views to 10-15 on another entry - is that an overload? Whereby in order to tell a compelling story, multiple charts need to be used just to 'add' more data visualisations, and have something to support the text?

Clutter is still the devil.
If 'dashboard design' IS a skill, then where does the ability to hold a narrative and not overly clutter come in? This brevity and ability to be concise is perhaps even more important than having a beautiful, colourful visualisation.

Perhaps this is an indicator to uphold a strong narrative the 'one screen' which I read - but my question is - is adding this level of 'scrollytelling' taking 'one screen' a bit too far? Should each segment of analysis be done using tabs instead? (Not story points, as I think this feature needs a bit of work - but that's for another post!)

The right tool?

One of the most brilliant things about Tableau is the simplicity to create something - it's quick, intuitive & once you 'get' the software, there's a lot more depth you're able to bring in. However, I've seen a few members of the Tableau community start dabbling with d3.js - which I think does scrollytelling correctly; the marriage of text and visualisation works well when it's integrated in the user experience - the first chart on this visualisation is a case & point of that.

So what's next?

I absolutely don't think that this trend will slow down at all.

The process of designing something 'beautiful' is inherently linked with long-form visualisations, because of the real-estate given to add a whole host of elements. However, this also means the shift away from some of what I believe to be core Tableau concepts - the championing of data visualisation, the ability to have some - but not total - design control.

What I ultimately want to do here is start a conversation. Given the statements and assertions I've made above, I want to hear more about what the community think - particularly on where this is going & whether it's something which should be discussed more.

As always, I welcome all thoughts, feedback & challenges - Comment, or get in touch! I would love to chase this down the rabbit hole further.