Data Science & Sports An overview Delft, April 7 th , 2016 - - PowerPoint PPT Presentation

Data Science & Sports An overview

Delft, April 7th, 2016

Kamiel Maase, Netherlands Olympic Committee

Sport Science & Innovation Program

MAIN GOAL: performance enhancement (medals!)

Support athletes, coaches, and their staff. By:

1. Distributing/ implementing knowledge

 Factsheets, protocols  Seminars  Sport science information center and helpdesk (Topsport Topics)

2. Hiring experts

 Sport dietitians  transfer of expertise, dietary logs, anthropometry  Embedded scientists  daily support, measurements

3. Infrastructure

 Field labs, climate room  testing, training  Embedded scientists platform & network

4. Research and innovation

 Research Program Sport  Sportinnovator Program, own projects  often with companies  Eat2Move (research and innovation on nutrition)

DATA IN SPORTS, A POSSIBLE CLASSIFICATION

But, there is more

Picture: LinkedIn.com

Types of data, a possible classification

• Recreational sports (‘organized’): monitors: participation data,

club members, KISS (kengetallen sportparticipatie);

• Recreational sports (‘not organized’): e.g. GPS data, social

data and city characteristics  quantified self, large datasets;

• Elite sports 1: sports intelligence (competition results,

performance progression, performance outlook (‘funnels’), benchmarking  tool for investment decisions;

• Elite sports 2: deliberate measurements

 Athlete measurements (physical, mental…)  External/environmental measurements  Competition analysis (technical, tactical)  Involves technology like sensors, video… Zoom in…

Examples (elite sports 2)

Regular measurements

• Heart rate, speed, contact times, position, power, personal logs;
• Technical and tactical parameters (“tagging/ scouting”);
• Anthropometry

Reasons to measure and

record

• Steering of training

 Direct feedback, learning

• Match/race preparation and

 Match/race analysis prior to next round  Protests

OPPORTUNITIES & CHALLENGES

The era of (big) data science

Opportunities

Larger data sets (elite sports: pooling of data, mass sports: Q-self)

• (Potentially) better analyses
• Danger: haystack

Combine data

• From one-dimensional to multi-dimensional (multi-disciplinary)
• Expected and unexpected correlations – data science

Computing power

• Power of the Crowd

Research programs

• Get the max of of your data:

 Better performance  Vital society  Understanding the value of sports

Picture: LinkedIn.com

Challenges

Quantity and quality of data

• Most sports data sets are not ‘big’
• However: mixed nature (from research projects to regular

measurements)

• FAIR data: findable, accessible, interoperable, reusable

 data formats, filing, and analysis Rules and regulations

• Privacy issues
• Competitive edge (data of elite athletes)
• Sports Data Valley

 Learn from existing initiatives!

This is just the beginning…