Why Vizioneer?

My photo
Atlanta, Georgia, United States
The "Vizioneer" comes from mashing two words that have shaped my world for most of my adult life - Engineer and [data] Visualizations (or Vizes to those who know what's up). Graduating from first from Oglethorpe University in Atlanta, followed by Georgia Tech with my Bachelors and Masters in Civil Engineering, all of which taught me to think through anything and everything - problem solving, "engineering" solutions, teaching to the "ah ha" moments - is what I love to do. In 2010 that investigative, engineering mindset intersected a job change and a plunge into the world of Data Analysis. In the search for the next great thing I stumbled on to a data visualization and dashboarding product called Tableau software and things just took off. So now I guess you could call me that engineer with the sweet data visualizations - or just "The Vizioneer" :)

In 2013, I joined the incredible team at Slalom, focusing on Tableau and it's been an amazing experience. Recently in 2014, I was honored and humbled to receive Tableau's highest recognition of being named a Tableau Zen Master. Follow along to see what happens next :)

Sunday, April 27, 2014

Day 27: Blending without Common Dimensions

The first time I met Dan Montgomery was about a month in to my Slalom experience.  He was visiting the southland from his home at Slalom Chicago.  He walked in and we immediately started chatting about Tableau.  It was like we were long lost friends who spoke our own language and we were showing off our work from different projects like it was first grade show and tell.  Since then we've both stepped up our game, both winning Viz of the Day twice and taking our Tableau skills to the next level.  So today, it's my pleasure to share this really insightful tip from Dan.  I give you Day 27 - Blending without Common Dimensions:




Sometimes when you're working with data, there is a need to compare datasets that don't really have anything in common.  This issue came up with my Sports Viz Show Me The Money: I had a dataset that contained the top 100 earning athletes in 2013 and I wanted to show what their pay could buy for them at an individual item level with out creating an unnecessary Cartesian product.














Athlete data: a single row per athlete

As Gru would say: Light bulb!  If I create a data set containing a row for each purchasable item repeated the maximum amount of times it would appear (aka the number of times the top earning athlete, in this case Tiger Woods, would be able to purchase it), then I could write a calculated field that blended the two sources and did my filtering for me.


















Stuff to buy data: note there is no column mentioning athlete

Doing this gave me two data sets: one that was 100 rows deep and the other just over 5,000 rows deep.


Next, after bringing in both data sets into Tableau, I created a cross dataset calculated field on the 'Stuff to Purchase' dataset that would allow me to filter out when the number of items exceeded the athletes pay.















running_sum(sum([Cost]))<=sum([Forbes 100 Athletes].[Pay $])

This way when I add the purchase able items at a data level, it will continue to referenced the accumulated value of the purchases and restrict them to only total that is less than the salary of the specified athlete.

Finally, I added
- RowID into the details
- 'How Many' true false filter to the filter section
    - IMPORTANT: Change the 'Compute Using' to use RowID
- Add the purchase category and item to the filter section
- Added the object's description to shape
- Filtered it for the specified athlete (from the secondary source).
   - This ultimately ended up being a quick filter that allowed me to select different athletes


Now I had my sheet that showed each of the items that the person could buy.





























If I had joined them ahead of time, I would have had a Cartesian product of over 160,000 rows!  This not only saved me data space, but also cut down on the data prep I needed to get dashboard built.  

So the answer all along (to the question that was never asked) was yes, it will blend...even if there are no common dimensions.  Let me know if you have any other questions and feel free to comment.  Thank you!

Thanks for following along!
Nelson

No comments:

Post a Comment