How the EXCEL?! / Tableau

The Magic of Tableau: Small Multiples and Faceting by Design

Every now and then, one of my clients will ask a question that sends me down a rabbit hole.  I spend a lot of time explaining, sometimes defending, and other times wondering right alongside them.

I started using Tableau around 2010.  My area didn’t have a user group and I had to really break my head to grasp what in the world was happening when I started.  Then, at some point, it started to make sense.

Here’s what I’ve come to learn – and yes, I could be way off on half of this – so take it with a grain of salt.  Or, maybe enough to clean my sidewalk…I’ll make the hot chocolate if you clear out the snow.

What started life as a small multiple machine has grown into a BI platform that is changing how people think about dashboards.

Chris Stolte did his thesis around exploring information visually and one of the go-to charts was, you guessed it, the small multiple.  Think about it.  There’s lots of patterns you can see this way.  You can get (think college data and Excel) everything in one view (again, think size and scope here – this is 2003 when cubes owned the world).  You can see, you can explore, you can ponder and wonder and dig under and answer questions that previously didn’t get exposed.

The way we’ve historically been taught to analyze data – include ALL the filters

Small multiples really set the stage for Tableau.  It focused on dimensions, not measures, for a change.  Everything in the past – Excel, SPSS, you name it – placed the most emphasis on the measure, or “The Number”.  For example, in Excel, you spend a ton of time writing IF or GETPIVOTDATA to expose The Number.  The Number is mythical: it is the answer to life, the universe, and everything.  Now if only we knew the question…

Tableau takes care of the number.  So, instead of the quest for the number, you as a Tableau analyst can explore the words, IDs, and other facets of your business, or The Dimension.  Take that in for a moment.

This is really a map of where I go in a day…

I didn’t get it for awhile, probably more time than I should admit.  I made dashboards for eons around The Number.  The Number is mythical, after all.  We will harpoon this Number!

Small multiples set the stage for dashboard design in Tableau.

Small multiples present facets of the same data through different lenses.  At first, Tableau didn’t have a way to make dashboards.  Really, no joke, it was just charts.  Then, at some point, they said, hey, let’s put multiple charts together.  And they did.  Then they made it so you could use charts to interact.  If you can use a chart to interact, it’s like faceting everything through a small multiple.  You’re using dimensions, not numbers, to explore.

In the time before, we had to use drop down filters.  You know these, you, I, and anyone on the planet working with data has seen and made a drop down.  We still use these.  But, put a few strategic charts, and you can drop how many you use and – with just a click, hover, or tap – see your data as never before.

And, here’s the real shocker…

The way Tableau is built, using charts for filters is more efficient.  And, as you learn to use this technique, it empowers you to give the users unprecedented ways to understand their data.  How else would I realize the only room in my house that’s consistently comfortable is my bedroom?  (Note to self: more ‘working from bed’ days.)

Faceting supports simplicity.

To do small multiples by hand would feel really complex.  But, luckily, the computers draw them now.  While we may throw a lot of data on these, the point is not to look at each point one by one, but as a collective whole.  Are there patterns that we see?  Is one blob smaller and tighter than others?  Why is that one so all over the place?  (Hint: we had colds and close doors…)

As we present these patterns, our drill down allows people to spot trends in a way that conventional analysis, such as the hunt for The Number, has blocked.  Visualization is key, but so is faceting.  Using charts as a way to present the data at a high level (here’s some averages, the patterns by day, and overall time variances) and at a detailed level in context.

Happy vizzing and stay warm!  In Ohio, we put ice cream in our hot chocolate.  Xocorosa or dark chocolate peppermint?

Leave a Reply

Your email address will not be published. Required fields are marked *