Tips on working from home

Working from home has always been a desire of mine. Thankfully I have come to experience this dream for about two years. Despite this length of time, I, however, cannot deny some of the challenges…

Smartphone

独家优惠奖金 100% 高达 1 BTC + 180 免费旋转




Data Analysis of FIFA 2019

The dataset contains information of all 18207 players from the latest edition FIFA 19. There are 89 attributes including personal information like age, name, nationality, photo, club, wage, etc, and also player skill information like ball control, dribbling, crossing, finishing, GK skills and etc.

I will focus on the three questions below:

Q1: What’s the ratio of total wages/ total potential for clubs. Which clubs are the most economical ?

Q2: What’s the age distribution like? How is it related to the player’s overall rating?

Q3: How is a player’s skill set influence his potential? Can we predict a player’s potential based on his skill set?

First I would like to look at the data at the clubs’ level. There is a total of 651 clubs collected and on average 27.6 players for each club. So which of the clubs are richest among them and which are more cost-effective?

To find answers to the first question, we can take a look at the wage/potential ratio for each club. The higher the ratio is, the more willingly a club spends money on high potential players.

On average, clubs spend €140 on wages for every potential of their players. So for a common player with 50 potential, his club will pay €7000 every year.

But it’s obvious that ‘Giant’ clubs are willing to pay much more on their players. Take Real Madrid for example, they have the highest ratio here even after Christiano Ronaldo transferred to Juventus. And this ‘money policy’ ensures them to stay very competitive: three times EUFA Champions League title in a row, king of Europe.

But when I take a look at the bottom of the ratio list, it’s a surprise that some club names are quite familiar to us. Like AEK Athens, Dynamo Kyiv, Shakhtar Donetsk, they all made impressive performance on UEFA Champions League. My guess is that the bad economic situation in their countries leads to low wages for players.

2. What’s the age distribution like? How is it related to the player’s overall rating?

From the above plot, we can see that most players are between 20–26 years old. And the players’ number starts to decrease after 26 years old and speed up after 30. The reason behind this could be that many young players didn’t get enough opportunities to prove themselves and give up their dream as a football player.

When a football player reaches their late 20s, they have gained enough experience and reaches the peak of their rating. The golden era of a football player’s career starts here and ends when his age reaches 35. At this age, his physical body condition drops quickly so as to average rating. As a result, they retire as a professional football player. But many of them stay active in the field of football, in the name of the coach, management role, and etc.

There are also quite a few numbers of players with age over 37, 38 years old. This is quite a surprise especially since their rating still can remain quite high. We should pay respect to their undying love for football.

3. How is a player’s skill set influence his potential? Can we predict a player’s potential based on his skill set?

At last, I wonder if there might factors that determine a player’s potential so I built a model to help us predict the player’s potential. I have used all of the player’s skill features as well as features like age. The model performs quite well since the r2 score is 0.87 on the test set.

As shown above, the potential of a player is mainly determined by his ball control, reaction, and age. Young players with excellent ball control and fast reactions tend to give us an outstanding performance in a football match.

In this blog post, we explored FIFA 2019 player dataset from Kaggle and got some interesting findings.

Add a comment

Related posts:

The customer games

Importance of customer mindset knowledge and the machine learning to acquire it to provide empathetic customer service via digital channels and chatbots during crisis. Loyalty matters.

Knightly Knews Episode 4

Eventful week to end 2021 with. As we look toward the New Year, I just wanted to say how grateful I am to each and everyone of you reading these articles. Let’s make 2022+ the year of the KNIGHT…

The Depth of a Beauty?

How do we see beauty at glance? Have you ever wondered why we think something is beautiful? I saw a lotus at the pond and I was amazed by it, followed by classifying it as beautiful, well in this…