Visualizing Flows

Sankey diagrams can display data in very interesting ways.


Normally when I hear the term process flow, I die a little inside. It just strikes me as management jargonβ€”a big pet peeve of mine.

Still, there are times in which the term flow just makes senseβ€”and not just on Six Feet Under.

Case in point: Sankey diagrams. Creating them is very simple with this tool and the case studies in The Visual Organization contain a few examples.

I grabbed the salaries of the 2017-2018 Cleveland Cavaliers from Basketball Reference:

Cavs' Salaries by Player in USD (2017-8)

Your Diagram TitleSalary β†’ LeBron James: 33,285,709Salary β†’ Kevin Love: 22,642,350Salary β†’ George Hill: 20,000,000Salary β†’ Tristan Thompson: 16,400,000Salary β†’ J.R. Smith: 13,760,000Salary β†’ Jordan Clarkson: 11,562,500Salary β†’ Kyle Korver: 7,000,000Salary β†’ Cedi Osman: 2,642,857Salary β†’ Rodney Hood: 2,386,864Salary β†’ Jose Calderon: 2,328,652Salary β†’ Jeff Green: 2,116,955Salary β†’ Ante Zizic: 1,645,200Salary β†’ Larry Nance: 1,471,382LeBron James: 33,285,709LeBron James: 33,285,709Salary: 137,242,469Salary: 137,242,469Kevin Love: 22,642,350Kevin Love: 22,642,350George Hill: 20,000,000George Hill: 20,000,000Tristan Thompson: 16,400,000Tristan Thompson: 16,400,000J.R. Smith: 13,760,000J.R. Smith: 13,760,000Jordan Clarkson: 11,562,500Jordan Clarkson: 11,562,500Kyle Korver: 7,000,000Kyle Korver: 7,000,000Cedi Osman: 2,642,857Cedi Osman: 2,642,857Rodney Hood: 2,386,864Rodney Hood: 2,386,864Jose Calderon: 2,328,652Jose Calderon: 2,328,652Jeff Green: 2,116,955Jeff Green: 2,116,955Ante Zizic: 1,645,200Ante Zizic: 1,645,200Larry Nance: 1,471,382Larry Nance: 1,471,382

Evidently, J.R. Smith’s salary doesn’t guarantee that he pays attention to the game’s score. (Couldn’t resist.)

Sure, I could have used a simple bar or line graph above. As I tell analytics capstone students, though, playing with new dataviz applications, tools, charts, datasources, and techniques is just plain fun. This mind-set is a major part of my teaching philosophy.

Marillion Studio Albums by Decade

I noodled around with SankeyMATIC and it’s fairly straightforward. Here’s a dataviz of Marillion’s remarkably consistent stream of albums by decade. I put my favorites of each decade in red. (Let the debate begin.)

Your Diagram TitleRadiation (1998) β†’ 90's: 1Script for a Jester's Tear (1983) β†’ 80's: 1Misplaced Childhood (1985) β†’ 80's: 1Clutching at Straws (1987) β†’ 80's: 1Seasons End (1989) β†’ 80's: 1Holidays in Eden (1991) β†’ 90's: 1Brave (1994) β†’ 90's: 1Afraid of Sunlight (1995) β†’ 90's: 1This Strange Engine (1997) β†’ 90's: 1Fugazi (1984) β†’ 80's: (1999) β†’ 90's: 1Anoraknophobia (2001) β†’ 00's: 1Marbles (2004) β†’ 00's: 1Somewhere Else (2007) β†’ 00's: 1Happiness Is the RoadΒ Vol. 1 (2008) β†’ 00's: 1Less Is More (2009) β†’ 00's: 1Sounds That Can't Be Made (2012) β†’ 10's: 1F E A R (2016) β†’ 10's: 1Script for a Jester's Tear (1983): 1Script for a Jester's Tear (1983): 180's: 580's: 5Fugazi (1984): 1Fugazi (1984): 1Misplaced Childhood (1985): 1Misplaced Childhood (1985): 1Clutching at Straws (1987): 1Clutching at Straws (1987): 1Seasons End (1989): 1Seasons End (1989): 1Holidays in Eden (1991): 1Holidays in Eden (1991): 190's: 690's: 6Brave (1994): 1Brave (1994): 1Afraid of Sunlight (1995): 1Afraid of Sunlight (1995): 1This Strange Engine (1997): 1This Strange Engine (1997): 1Radiation (1998): 1Radiation (1998): (1999): (1999): 1Anoraknophobia (2001): 1Anoraknophobia (2001): 100's: 500's: 5Marbles (2004): 1Marbles (2004): 1Somewhere Else (2007): 1Somewhere Else (2007): 1Happiness Is the RoadΒ Vol. 1 (2008): 1Happiness Is the RoadΒ Vol. 1 (2008): 1Less Is More (2009): 1Less Is More (2009): 1Sounds That Can't Be Made (2012): 1Sounds That Can't Be Made (2012): 110's: 210's: 2F E A R (2016): 1F E A R (2016): 1

If I were taking my analytics class as a college senior, then I’d go much deeper here.

For a much better example when this type of diagram makes sense, consider how where one Reddit user slept during over the past six months:

In case you are wondering, you can do the same thing in Tableau. Click here for a Sankey template.

Simon Says: Use new and interesting methods such as Sankey diagrams to represent data.

I can’t say it any better than Dr. Seuss:


What say you?




Filed Under

Enjoy this post? Click here to subscribe to this RSS feed or here to sign up for my bi-monthly newsletter.

Submit a Comment

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