People won’t use Tableau Prep to prep their data. My crazy idea of what they will do!

It was an exciting day. Tableau launched their first completely new product in years. If that wasn’t enough, the opening volley included new bundling models and pricing plans. I’m not going to get into what I think this all means for Tableau, there are way too many others jumping all over that.

If you’ve been following Project Maestro, well, you’ve been excited for Tableau to put their magic touch on the field of data preparation/cleansing. It doesn’t have all the features as some other products -- and I’m not sure it will ever need to match them feature-for-feature. But it’s going to help bring a lot of people into the space who may have been holding out because of price or software complexity. Those are both gone!

However, Tableau Prep still may not for everyone. I say that because it takes a certain type of thinking to break a problem down into individual steps to reach a final outcome. If your brain lines up that way, Prep works brilliantly for you.

Here’s my surprise idea. I think (almost) EVERYONE will use Prep to look at the data to understand it’s makeup — even if they don’t use any of the features to prep data for Tableau. They will quickly be able to know things like:

  • What elements are in the dimensions?
  • What is the range of each measure?
  • Does it look like there are too many nulls?
  • Is there data missing from a certain time?

Here's one of my favorite NYC Open Data sets in the first stage of 'prep.'  It's great to see that the highest number of violations aren't directly tied to food and a most of those related restaurants received an A grade.  It's safe to be hungry in New York.

You've probably been doing this type of analysis in Tableau Desktop.  In the same way Desktop increased the speed of analyzing data, Prep dramatically decreases the time to understand your data.

Prep makes it easy to catch the gremlins that get into your vizzes and make you crazy. I am hyped (pun intended) to see if this becomes a powerful, new component of how people start working with data.

Did you know 9 in 10 people just read social media posts and don't participate in the conversation.   Instead, be a '10 percenter' and leave a comment.  Or click the button ⏬ to read more.

  • OK. I’ll bite. Participate, I mean, rather than simply read and go away. But it’s really to ask a general data prep question rather than to comment on the commentary.

    I’ve never understood the general statement (which I hear a lot) that the structure of your underlying data determines what you can do with it visually in Tableau.

    More specifically, wide/short vs narrow/tall. Maybe my visualizations are too simple (80% bar charts), but I generally don’t see much difference in how the data comes in vs what I can do with it.

    Is restructuring the data more about getting a clean separation between dimensions and measures? Is it a matter of dividing what you filter on and what gets filtered? I’m not looking for an answer here — more a pointer to someone who has explained this subject well.

    My data tends to be very dimension-heavy. Lots of Likert scales and multiple-choice text responses. My measures tend to be counts and count distincts. And yes, I’ve read everything Steve Wexler has ever written. Still not quite there.

    • Lee Feinberg says:

      Steve is the go-to for Likert! I don’t categorize it as wide/short or narrow/tall, but flat (which does tend to mean narrow/tall).

      If you are connecting to a well-modeled data source, Tableau is going to work fine. Problems usually crop up with Excel files that are formatted like reports, not data sources.

      Tableau Desktop has features like data interpreter and pivoting that can help. Tableau Prep takes it to another level.

  • Ellen Simpson says:

    Here’s what I want to know… How does the Morris Park Bake Shop have evidence of live mice (row 3 in the NYC Open Data), but still get an A grade?? Yuck!

    Yes, I agree, Prep looks like a great new tool for the folks that need to look more closely at the data!!

  • Agree with your likely use case for Tableau Data Prep … basically inline basic profiling of the data similar to what you see in some of the other high-powered data-prep tools; Datawatch, Alteryx etc …

  • >