Data Exploration: Using Tableau Dashboards to See and Understand your data

I love Tableau like I love Giorgio Moroder.  One gave me music and both gave me insight.  Tableau tells me I can see and understand my data with a few clicks.  They’re not wrong.

What they don’t say about these clicks is I can’t go in expecting to find the click, at least not if I want to find anything new.  Giorgio Moroder used a click to set the pace, allowing him to match beats and create a metronome of sorts.  This click allowed him to create something new – music that has evolved into modern dance music.  Love it or hate it, it changed the game.

Tableau, like these clicks, gives me a pace where I can quickly rip through data.  It’s not a process I let people watch, as it really is like that kid on birthdays ripping through presents.  Trust me on this one.

I get a lot of data these days.  I rip through these patterns and I hope to give my clients insights based on what they request.  Sometimes, I find new information.  Other times, I confirm suspicions.  A few rare times, I make people angry – it happens, mostly because there’s some bad news somewhere.

As analysts, we’re often responding to a need.  We’re looking at some trends our boss requested around a new project.  Or we’re trying to figure out what went awry at 2 AM on Sunday. But to just dig into the company wells of data for no given reason other than to explore, that doesn’t often happen.  The rare few get this – if you do, enjoy it, it is a treat.

We often find something in any data.  But to really explore takes a mindset.  In the past, our ability to do this was limited.  We either didn’t have the data, didn’t have it in a low-enough level or usable format (oh, Pew Research Center, you are on my list…), or found ourselves locked out of the ability to do much beyond exporting a simple pivot.

When we connect to data – real, beautiful, in the wild, transactional style data, the world opens up.  The clouds part and we have these rays of sunshine called dimensions.  Dimensions are like little bits of vitamin D, while measures get to be like UV rays…there is a limit for everyone.

We can then use this data, preferably the parts with words or descriptions that mean something, to understand what is going on in the world.

Here’s a look:

This is a pattern we see fairly often in HR.  The curve may vary, depending on levels, but you can usually top level execs to make more than others (if you’re fired up and ready to send me the next Inc article on that CEO who gave up pay, don’t – this is a generality).

And another:

Same view, just turned over.  Let’s see how it builds!

Hey look – gender differences!  Can we look at this a bit differently?  Hmmm, I wonder…putting it on the same line seems a bit clearer.

The color calls out the women.  Naturally, I have to be contrary so it’s green and grey.  What if I add degree advancement?  Surely, that accounts for it!

Oh wait, there’s something here!  Let’s see if it’s related to departments.

Hey, this doesn’t look right!  Surely, something is going on….How about years of service?

The rabbit hole keeps getting deeper…let’s try a different way.

Let’s look at the tail end a bit closer…

Real admission here: I may not use any of these charts, but I did this in less than 5 minutes.  But now I know a few things from this (very fake) HR (wanna be) data set:

  1. I have a gender pay gap at the top.  Initially, I may be able to explain it with education or experience, but adding in degrees makes me realize women need more education in order to earn the same as a man.  I’ll want to look more into this, but this is pretty glaring so far.
  2. My sales team pay is all over the place, as well as in a few other areas.  I should probably look a bit more into this.  Maybe performance, age, or commission are factors?
  3. Especially at lower levels, it doesn’t look like I’m rewarding long term service.  In fact, there seems to be a penalty the longer you work with my (very fictitious) company.  But gender doesn’t seem to be an issue at this level.

Yes, I spent an absurd amount of time mocking up these trends, but they mirror a number of the findings that exploration reveals.  We often think we know what the data will tell us.  We’ve made up our mind and often make a lot of charts that (somewhat) support these assumptions.  Would I have found these if I was told to put together very specific charts or types of analyses?  Probably not.

And maybe that’s the secret: the value is in the journey.  Happy exploring!