1 Reply Latest reply on Mar 8, 2018 4:13 PM by Patrick Van Der Hyde

    My first Jump Plot in Tableau

    Annabelle Rincon



      I recently published a new viz about the Tour de France after several weeks working on the data very late at night (thank you wikipedia) and the vizualisations. https://goo.gl/z1HNxF ( feedback welcomed).


      In one of the viz, I wanted to compared the top 10 riders across the different Tour editions, and especially focus the reader's attention on the outliers. A box and whiskers seems correct to me, but I wasn't totally convinced to be able to distinguish between the editions. At that time, I was watching the Tableau Conference videos, which are my favorite tool to increase my Tableau and design skills.


      My curiosity crossed path with the excellent video of two giant Tableau Zen Masters  Adam McCann and Christopher DeMartini about New Advanced Chart types : Jedi Charts: Creating New Advanced Chart Types - YouTube

      This is an excellent video, funny and very instructive, that I highly recommend.


      During this session, Chris DeMartini and Adam McCann presented a new graph ( created by Chris DeMartini and Tom Vanbuskirk), that I have never heard before : the Jump Plot. According to them a Jump Plot is very useful to identify outliers in repeated event sequences and/ or if you want to identify  root cause of process delay. A Jump Plot is a good way to visualize event data ( ex supply chain or manufacturing workflow), because it allows you to verify if your process follows all the steps, and the time between the different steps.

      The steps ( checkpoints) are connected by a hop with a beautiful bezier curve, and the hop height is determined by the time between two steps.


      Capture d’écran 2018-02-24 à 15.02.45.png

      example : from Jump Plot | A new way to visualize event data



      This video gives me the idea to use this representation in my new dashboard; in this case :


      - each checkpoint represents the position of a rider on a specific Tour edition

      - each line an edition

      - the hop p represents the time between the 1st rider and the rider, who arrived in the p position.


      Capture d’écran 2018-02-23 à 22.23.57.png



      To do it, you need to have a specific data structure ( cf references at the end of this post), in my case I have joined two tables to achieve it, because it will duplicate the rows, as needed.


      Capture d’écran 2018-02-24 à 15.20.27.png



         classement 10              

      Capture d’écran 2018-02-24 à 15.29.09.png


      path order

      Capture d’écran 2018-02-24 à 15.29.39.png


      Once you have done it, you have to create some fancy calculations ( Chris DeMartini's website explains all the steps perfectly).


      Finally, I give some flexibility to my viz:


      - as a decade filter, in order to allow the user to focus his attention to a certain group of editions, due to technological improvements it is not fair to compare the riders performance in 1918 and in 2016 ( for instance the state of the road in France after World War II was a disaster);


      - a highlighter, so that you can easily identify the Edition when you click


      - a second viz at the bottom, that give you the name of the top 10 riders for the clicked edition.


      Capture d’écran 2018-02-24 à 15.33.41.png


      I hope you like it.





      Absolute References ( they speak better about it than myself ):


      How to make Jump Plots in Tableau | Tableau Public

      Time to Get Hopping with Jump Plot by Chris DeMartini and Tom VanBuskirk — DataBlick

      What is it? | Jump Plot