While looking for interesting datasets on Kaggle, I came across this amazing dataset about the 2019 AO Men’s Final Game. The dataset’s creator has meticulously collected comprehensive data regarding the positional, temporal, and stroke information from the final match featuring Rafael Nadal and Novak Djokovic. It’s a dataset ripe for visualisation in Tableau, so I figured I might as well take a crack at it.

As a warm up, I chose to examine the serve placement executed by both players throughout the entire match. The author of the dataset has kindly provided the code for the dimensions of the court, as well as the positions where the successful serves bounced. I created the background image for the court in Figma based on the dimensions used by the author in his data collection. Following some data cleansing to ensure all coordinates are reflected on a single half of the court, here’s the visualization I arrived at.

Right away, I noticed that something was off. Given that my visualisation should only include successful serve placements, all the serve bounces should have landed within the service box. After reconfirming my court dimensions (and seizing this chance to revisit some of the rallies on YouTube), I decided to backtrack and little and visualise the serve placements on the full court.

From this perspective, we can clearly see where parallax errors may have occurred in the data gathering process. The author did note in his documentation that “the positions are shifted a bit to the left” due to shifts in camera perspective (you can guess which way the court was oriented relative to the camera). From this visualisation, we can also see that the recorded location of shots on the far end of the court are also marked as farther than their actual location. It’s a bit unfortunate but it’s still plenty to work with, and kudos to the author’s efforts in the absence of publicly available Hawk-Eye data.

Now that my appetite for this dataset is whet, I’m ready to dig deeper to see what other visualisation opportunities this dataset has to offer.

The Data School
Author: The Data School