-- Mohammed Ayub. References Sports Data Mining- Springer by Robert P. Schumaker, Osama K. Solieman, Hsinchun Chen

Data Mining Applied to Sports

-- Mohammed Ayub

References• Sports Data Mining- Springer by Robert P. Schumaker, Osama K.

Solieman, Hsinchun Chen• http://dataminingsoccer.com/en/data-download/ - Historical Soccer

Dataset. • http://www.sloansportsconference.com – MIT Slogan Sports Analytics

Conference • https://www.acmilan.com/es/club/milan_lab - AC Milan Lab• http://www.clubmilan.net/?cat=2&subcat=2&details=6 – AC Milan

Lab Working• http

://www.zdnet.com/ac-milan-the-high-tech-giants-of-european-football-3040145126/ - January 2004, ZD Net

• https://www.youtube.com/watch?v=F8TtbVpZVYE&hd=1 - Demonstrating the working of Milan Lab http://users.cis.fiu.edu/~chens/PDF/ICDE05.pdf - Paper on SoccerQ video retrival software.

Why Focus on Sports Industry..??

• Vast amount of data collected over time for Players, Teams, Club.

• Very less Preprocessing required.

• Maintain Competitive environment.

• Multi Billion dollar Industry.

http://www.wildkingdumb.com/2013/09/blog-post.html

Evolution of Relationship between Sports and Sports Data

No Relationship

Domain Experts + Gut Feeling

Domain Experts + Historical Data

Use of Statistics for Decision Making

Use of Data Mining for Decision Making

Hmm

http://blog.zopim.com/2013/11/28/evolution-sale/

A Few Success Stories of Data Mining in Sports

• Billy Beane’s Oakland Athletics (A’s) – nearly defeats New york Yankees in 2001 –using “Sabermetrics”

• Boston Red sox Wining two World Champions(2004 and 2007) after a 86 year gap.

• Ukraine’s Kiev Dynamo wins Union of European Football Associations (UEFA) Cup in 1975 and 1986.

• AC Milan reduces player injuries.

Problem with Statistics

• In Baseball – metrics such as Batting Average, Earned Run Average (ERA).– Refined Formula’s by James– Runs Created = ((Hits + Walks) * ΣBases) / (At-Bats + Walks)– ERA = (Earned Runs Allowed * 9) / IP

• In American Football –No of Receptions, Yards per carry.

• Basket ball – Rebound Statistic, Field goals Percentage

Prediction Player Injury to increase Player Downtime

• AC Milan’s success story behind a healthy team.

Jean-Pierre Meersseman AC Milan Lab

www.acmilan.comhttp://www.theguardian.com/football/2013/feb/16/milan-lab-premier-league

• Computer Associates (Brightstor, CleverPath and eTrust software) developed an analysis software.

• Supervised Classification method- Neural Networks is used to predict the Injuries.

• Unisys system - gathers physical and mental player state with help of AMD Hardware.

• Information is fed into an analysis software developed by CA and analyzed using PAS(Predictive analysis Server).

• Data set consists of Injury and Recovery History, Diet, Performance, Biochemical and Skeletal statistics.

• 91% reduction in player injury rates.

• Reduced 41 muscle injuries to just 2-3 for three consecutive years.

• Helps in player selection, organization, and player trade during transfer season.

https://www.acmilan.com/it/news/show_photo/923/30734

THANK YOU ..!!!

Questions..???