3 Replies Latest reply on Nov 29, 2016 8:09 PM by Eric Waschak

    Customer Journey Visualization - Slopes? Polygons? Other?

    Eric Waschak



      I'm attempting to visualize a customer acquisition process. I work in higher education, so I'm attempting to visualize prospective students moving through the enrollment process, but I'd imagine the principles apply widely across many industries.


      My data set has ID numbers with various dimensions (e.g. ethnicity, high school origin, geographical location, Anticipated Term Start) and measures (e.g. GPA, SAT). It also includes the date a record reached a particular point in the enrollment funnel (e.g. "Application" with a datetime value) and the term start date. Something like this:


      ID #Inquiry Date
      Application DateEnrolled DateAnticipated Start TermAnticipated Term Start Date
      100018/1/15 1:57:39 PM8/5/15 3:40:01 PM 8/7/16 4:35:16 PMFall 20168/31/16
      100028/4/15 2:50:31 PM 8/7/15 10:57:21 AM 8/12/15 9:00:00 AMFall 20159/2/15
      100038/3/15 4:57:39 PM 8/5/15 7:59:13 PM8/6/15 8:00:15 PMFall 20159/2/15


      I've used DATEDIFF() to calculate how many days before the term start these actions took place. I've found it is far more meaningful to visualize changes in the funnel with reference to the term start, since that normalizes the data. (Terms don't start on the same day each year, it's impossible to compare Fall with Spring, etc.)


      I'd like to visualize each new student's journey to enrollment for a particular term. This format disaggregates customers while allowing trends to emerge. I'll add other exploratory functionality later. Here's a sketch of what I'm attempting:


      Is there a way to create a chart like this without reshaping the underlying data? I've looked into slope graphs, but they appear to all rely on a single measure changing with respect to a date dimension. That won't work for this visualization because there's no set timeline for a student to take these actions, but they all occur in the same order. I've seen some very clever visualizations using polygons, but those I've seen require shaping the data source in ways that would break the other visualizations in my workbook.


      I've attached a dummy workbook that includes the visualizations I've already created (with randomized data).