How to do Sports Analytics on Power Bi?

I have always been passionate about football or ‘soccer’ as popularly known in America. Watching the English Premier League, and then selecting the most suitable players week in week out for my “Fantasy League” team has been a part of my routine in the last few years. However, following every action that takes place in the league sometimes get difficult and I would always miss out on some or the other key stats (especially of the players from the bottom of the league teams).

This is when I thought of combining one passion with the other, that is Football with #analytics

I began my search for data of players’ statistics and then found the desired data on https://www.kickest.it. By using ‘BeautifulSoup’ and ‘Pandas’ packages in #python , I could scrape the website and extract the data I needed in .csv file format.

I imported this .csv file to my Power Bi Desktop and then began my analysis.

The ‘Transform Data’ feature of Power Bi made it very easy for me to clean the file so I could consume this data in the format I wanted. I could use this to build new metrics like ‘Goals per Shot’ or ‘Clean Sheets per Appearance’ etc. with no hassles using the ready-to-use DAX functions.

I started designing my dashboard mainly in 4 sections. Firstly, an overall analysis of ‘21-’22 EPL season displaying the most important stats from the league like ‘Top Goal Scorer’, ‘Most Clean Sheets’, ‘Most Assists’ etc. I added drop downs to analyze the same data by Team or by Players’ Field Position. Further, I designed 3 more sections for the 3 main field positions that are – Attacking players, Midfield and Defenders. I added metrics in each section based on the role of these field positions in the game of football. For example, I would like to compare Attackers based on Goals they score or assists they provide, whereas I would analyze Defender based on the clean sheets kept or tackles won.

Easy to plot and aesthetically pleasing #visualizations offered by #powerbi made the process of analysis a walk in the park. I could easily plot metrics like ‘% of Goals scored by each team’ and compare these against other teams in a Donut Chart. Heat Maps to find out the players receiving the most number of yellow cards. A comparison of ‘Most Goal Contributions by Defenders’, along with ‘Goal Contribution per Appearance’ ratio etc. By using simple ‘button’ bars, I could compare the Attackers of bottom of the league teams based on the number of goals they score or number of shots they take in every match.

This is not all. The most exciting feature is the #insights visualization. Power Bi automatically analyzes the data on each page and displays the most useful information in the ‘Insights’ section based on the selection done in filters. For example, when I compare bottom of the league teams’ attackers Zaha, Toney, Pukki and Richarlison, ‘Insights’ section displays “ At 14, W. Zaha had the highest Goals and was 40% higher than Richarlison, which had the lowest Goals at 10.” Or amongst all midfielders - “ K. De Bruyne accounted for 13.27% of Goals Scored”, isn’t this cool?

This is how I built my first #sportsanalytics dashboard on Power Bi by using any spare time that I had over 3-4 weekends. Do have a look at few of the snapshots here and share your thoughts.

Note: With this post I do not intend to promote betting or any illegal actions related to betting.

The EPL logo image used in dashboard is copyright of English Premier League and is used in the dashboard only for learning purpose only.

The Source of Data: https://www.kickest.it/

Tools Used: Python, Microsoft Power Bi

#analytics #sportsanalytics #powerbi #python #webscraping #datainsights #businessintelligence