002 | ATP 2023 Summary | A Tableau Dashboard
Tableau is an amazing tool for data visualisation. It is intuitive in it's understanding of different data sets, contains a number of different templates that can be customised and incorporated into dashboards and Tableau Public is actually available to anyone, for free.
To show off some Tableau's capability, I wanted to identify a dataset that was structured but disparate, easily obtainable but most importantly would be of interest to a relevant audience.
---
OBJECTIVE
The main objective was to demonstrate how the Top10 Male tennis players in the world had earned their ATP Ranking points throughout the year. I also had a few smaller hypotheses I wanted to test:
Hypothesis 1. Novak Djokovic had earned his deserved No 1. ranking through performance in the Grand Slams vs other qualifying ATP tournaments.
Hypothesis 2. Unlike in 2022 there had been no breakout stars accelerating through the rankings in 2023 and entering the Top10.
---
DATA PREPARATION
I obtained data about each players performance directly from the ATP Website and consolidated this into a spreadsheet. I then cleaned the data by ensuring that there were no duplicates, evaluating any outliers and removing any irrelevant data.
1. Not every player competes in every event. This led to null values when transforming data using pivot tables and created challenges for generating some of the calculations.
2. Some events occur simultaneously. This led to unexpected duplication of the date field. To solve for this I wrote a formula to identify duplicate dates and concatenate tournaments based on having the same date.
3. Date field recorded only as text. A common problem when importing data into Google Sheets is that it can only recognise dates as text and even resists when you try and convert the format. To overcome this, I use the a combination of left(), mid() and right() formula to rebuild the data in the
=right(F13,2)&"-"&mid(F13,6,2)&"-"&left(F13,4)
---
ADDING CALCULATIONS
The key calculation required from the data was the cumulative ATP Points throughout the year. With all the data in one table, adding this data required a small formula to add player points after each tournament (Column G) but importantly reset for each player (Column A)
---
THE TABLEAU DASHBOARD
Tableau has simple connections to multiple different file types meaning importing the data was relatively straighforward. I created a variety of different visuals across different sheets and merged them into the dashboard here. A slightly squashed version can be seen below:
---
CONCLUSIONS
Hypothesis 1: Djokovic earned the majority of this ATP Points through performance in the grand slams.
Djokovic scored 72% (7200) of his overall points in the year in the Grand Slams. In what was an almost perfect year, the Serb won the Australian Open, US Open, French Open and was edged in 5 sets by Alcaraz in the final at Wimbledon. Beyond the Grand Slams he entered 8 other tournaments, almost half the average of the other players. However, his performance at slams and the ATP Finals alone would not quite have been enough to finish the year in no.1 position as Alcaraz's 8,845 points picked up across 16 tournaments would have trumped the 8,500 Djokovic scored from these events.
Hypothesis 2: Limited movement within the Top 10 Players in the world throughout 2023.
Whereas 2022 saw the emergence of young starts Carlos Alcaraz, rising from 32 to no.1 in the world and Holger Rune from 111 to 11, 2023 saw a much more stable Top10. The biggest riser throughout the year was perhaps the player who had the strongest end to the year winning the Davis Cup for Italy. Jannik Sinner began 2023 ranked 15 in the world and finished ranked no.4. Lost to the Top10 were Rafael Nadal through injury and Casper Ruud and Felix Auger-Aliassime who struggled for form in the latter part of the year.
---
KEY LEARNINGS
This was the second Tableau dashboard I have ever built and quite a bit of trial and error was required to reach the eventual output. As well as the conclusions above, I wanted to document some of the key learnings so that I and others may avoid such mistakes in the future.
1. Tableau does not always play nicely with crosstab data. There are some easy solutions available to solve this but I now know not to invest extra time getting data into a crosstab format for Tableau.
2. Refreshing base data does not always work. If I added data to my Google sheets and refreshed in Tableau it did not always recognise new data. I found it easier to create a new sheet, import it and link them.
3. Dashboards do not resize well for different screen sizes. I understand that this is particularly the case when you are using floating elements vs tiled. I have learnt that it's best to design for a smaller screen so that it can be upscaled.
