Sports Analytics
Innovation Summit
Winning in Sports Through Performance
Analysis & Data Analytics
September 12 & 13, Boston 2013
Who went?
Coaches, General Managers, Sports Analysts, Technologists
Key questions
How do you engage Coaches?
How do you create value based on the Data?
Two tracks
On Field Analytics
Off Field Analytics
The pursuit of
innovation in the English
Premier League
Tony Strudwick
Head of Sports Science at Manchester United
The pursuit of
innovation in the English
Premier League
Tony Strudwick
Head of Sports Science at Manchester United
Psychology of Sport
Analytics may not capture a
player's main strengths or
weaknesses
For example ...
One players greatest skill may be
to motivate other players.
If players are not physically or
mentally ready to perform then
data is a waste of time.
Other Problems ...
In the last 20 years, sport science
has been oversubscribed yet has
underdelivered.
Most coaches feel sport science
brings no value to their team.
The problem is perceived as the inability of
data practitioners to communicate
actionable metrics.
Analytics do not always explain human
psychological principles because ...
Humans are not rational
Humans are risk adverse
Under pressure humans will fail
Analytics must drive decisions
and actions or else they're
worthless.
Need more graphical representations, not
excel spreadsheets
Emphasis on real time apps, real time data
& analysis, real time decisions
Catapult Team Tracking System
Decisions that are now well
supported by analytics ...
Managing training of new players
Analytics & reports for chief execs
Display relevant relationships
between these variables:
Age Group
Position
Ave body Mass
Cumulative minutes trained
Example
49% of squad is 29+ years old
Higher number of injuries coming from
this group
During November, December, January
intensity training goes down, injury
goes up
Goals of social media to
analysis
Using

search for the next
Olympic team
Remain injury free

Increase player availability and individual
performance
Troy Flanagan
Maintain high performance over 45 games
Performance Director, US Ski & Snowbaord Assn.
per season, 4 games per week
Using social media to
search for the next
Questions?
Olympic team
Troy Flanagan
Performance Director, US Ski & Snowbaord Assn.
Using social media to
search for the next
Olympic team
Troy Flanagan
Performance Director, US Ski & Snowbaord Assn.
Goal of program is to transfer ex-gymnasts
into aerial skiing for the 2018 Olympics
3 years to reach the podium ...
The US Ski Team created a Facebook app
through create.it for finding talent
Kids submit their best tricks to win an
invitation to tryout camp
If theres not a tangible reward people
won't participate
Using Visual Analytics in
Performance Analysis
Questions?
Kirk Goldsbery
Visiting scholar at Harvard, now at ESPN
Using Visual Analytics in
Performance Analysis
Kirk Goldsbery
Visiting scholar at Harvard, now at ESPN
What are Analytics?
Analytics are Reasoning Artifacts …
things we use to make decisions.
New Data, New Analytics, Same reasoning
Maps
Maps show spatial structure and patterns
Maps provoke spatial reasoning
Maps work for all sports
We’re visual creatures
and when we see something attractive we want to
consume it

It takes time
to make something that people want to consume.

If you were to ask Faulkner how he writes …
he doesn't just write, he considers how to frame the
story first
How do you harness spatial science?
Sports are spatial
Sports are visual
Analytics are not spatial or visual
Spatial Analysis
Visualize patterns and quantify information

Visual Analytics
Translate raw game data into useful information
Example
All LeBron James shots for the last 5 seasons
Spatial map of shooting patterns
Good for engaging the athlete
Good for finding players that are similar or different
Visualizations can handle big data
As strategic devices
As communicative devices
Two different approaches to visualizations
Exploratory
Confirmatory
Depends on audience
Sensors
lead to quantitative spatial research questions
However, provoking spatial reasoning may lead to
more questions than it answers
Find a question to try to answer and attack it
Staying Connected:
The Rise in Fitness Data
Questions?
Chris Glode
GM, MapMyFitness
Staying Connected:
The Rise in Fitness Data
Chris Glode
GM, MapMyFitness
MapMyRun
Team of 100 split between Austin & Denver
Connected Fitness - social, fun, simple, effective,
and rewarding
Users
40% Aspirational

45% Recreational

15% Fanatical
160 million workouts logged in 2013
Team working toward less friction in the
app experience
Growth driven by
Smart phones
Wireless technology & reduced friction seamless data download
Cloud computing
Wearables… reduction in hardware costs
Obesity epidemic in US
People use MapMyRun to “outsource
their willpower”
Friends in the system keep other users
more active via notifications
Techniques for engagement:
Games: people keep coming back for competition
Games are considered a “jedi mind trick” by
MapMyRun, effectively manipulating users to return
Route art: motivated users who were otherwise
uninterested in social
Practical applications of fitness data
Corporate wellness
Trainer driven programming, tailored to the individual
based on real, recent and new fitness data. Total
accountability
Opportunity via tons of information mined
on the geospatial front
Example: advertising to women along
certain routes, etc.
Future
Woven wearables
Advanced activity detection
Ubiquity of incentives to track fitness
iOS7 - Support for passive all day activity tracking in
background when app is inactive
Using GIS to study
Spatio-Temporal
Questions?
Patterns in Sport
Damien Demaj
Geospatial Product Engineer at Esri
Using GIS to study
Spatio-Temporal
Patterns in Sport
Damien Demaj
Geospatial Product Engineer at Esri
Space and time go hand in hand in sport
Mapping a tennis match
Example
Who: Federer v Murray
Data: 1706 spatial points
3D GIS & streaming video
The serve: the most important shot
Speed & spin: the most important metrics
Variation is key: mapping unpredictability is
important
Approach
1. K Means algorithm - looks for natural clusters in the
data, balls that are close in space but also have a
similar attribute
2. Create euclidean lines & calculate large mean distance
3. Tag the most important points in the match
4. Add a feature overlay – Pseudo Realism – putting
players back in their environment
Smart Soccer
Qaisar Hassonjee

Questions?
VP Innovation, adidas

Nelson Rodriquez
MLS EVP Competition & Game Operations
Smart Soccer
Qaisar Hassonjee
VP Innovation, adidas

Nelson Rodriquez
MLS EVP Competition & Game Operations
How is technology enabling
sports analytics?
Wearables
Multifunctions, always connected, smart/
aware devices that measure me
Enabled by advances in sensor
technology, algorithms, data science
Everyone from startups to established
brands are developing tools
What are you looking for? What kind of
sensors to develop, ease of use
Lots of fear in the industry... but lots of
copycats once something works
The adidas Team System
Open platform… more sensors over time can be
added to the platform
Measures Heart rate, Speed, Distance, Location,
Acceleration
100 shirts
30 pods
4 ipads
19 MLS clubs are now using the
system - bell curve of adoption
iPad App
Most people don’t look at all 20-30 parameters,
just the top 2-3
How is the information actionable?
How is pre-season trianing improving the fitness of
my athletes?
Results
Athletes appreciate and want to use the technology
Tool can extend careers and improve performance
Injuries are down 2% this year
Future
Drop the tech into the academy programs & build a
national data set of kids & analytics using the system
Commercial opportunities, super fan stuff, fantasy
teams, etc.
USA Volleyball
Questions?
Anton Willert
Technical Coordinator/Tem Manager
US Men’s National Team
Thanks for visualizing

Sport Analytics Innovation Summit

  • 1.
    Sports Analytics Innovation Summit Winningin Sports Through Performance Analysis & Data Analytics September 12 & 13, Boston 2013
  • 2.
    Who went? Coaches, GeneralManagers, Sports Analysts, Technologists
  • 3.
    Key questions How doyou engage Coaches? How do you create value based on the Data?
  • 4.
    Two tracks On FieldAnalytics Off Field Analytics
  • 6.
    The pursuit of innovationin the English Premier League Tony Strudwick Head of Sports Science at Manchester United
  • 7.
    The pursuit of innovationin the English Premier League Tony Strudwick Head of Sports Science at Manchester United
  • 8.
  • 9.
    Analytics may notcapture a player's main strengths or weaknesses
  • 10.
    For example ... Oneplayers greatest skill may be to motivate other players. If players are not physically or mentally ready to perform then data is a waste of time.
  • 11.
  • 12.
    In the last20 years, sport science has been oversubscribed yet has underdelivered.
  • 13.
    Most coaches feelsport science brings no value to their team. The problem is perceived as the inability of data practitioners to communicate actionable metrics.
  • 14.
    Analytics do notalways explain human psychological principles because ... Humans are not rational Humans are risk adverse Under pressure humans will fail
  • 15.
    Analytics must drivedecisions and actions or else they're worthless. Need more graphical representations, not excel spreadsheets Emphasis on real time apps, real time data & analysis, real time decisions Catapult Team Tracking System
  • 17.
    Decisions that arenow well supported by analytics ... Managing training of new players Analytics & reports for chief execs
  • 18.
    Display relevant relationships betweenthese variables: Age Group Position Ave body Mass Cumulative minutes trained
  • 19.
    Example 49% of squadis 29+ years old Higher number of injuries coming from this group During November, December, January intensity training goes down, injury goes up
  • 20.
    Goals of socialmedia to analysis Using search for the next Olympic team Remain injury free Increase player availability and individual performance Troy Flanagan Maintain high performance over 45 games Performance Director, US Ski & Snowbaord Assn. per season, 4 games per week
  • 21.
    Using social mediato search for the next Questions? Olympic team Troy Flanagan Performance Director, US Ski & Snowbaord Assn.
  • 22.
    Using social mediato search for the next Olympic team Troy Flanagan Performance Director, US Ski & Snowbaord Assn.
  • 26.
    Goal of programis to transfer ex-gymnasts into aerial skiing for the 2018 Olympics 3 years to reach the podium ...
  • 27.
    The US SkiTeam created a Facebook app through create.it for finding talent
  • 28.
    Kids submit theirbest tricks to win an invitation to tryout camp If theres not a tangible reward people won't participate
  • 30.
    Using Visual Analyticsin Performance Analysis Questions? Kirk Goldsbery Visiting scholar at Harvard, now at ESPN
  • 31.
    Using Visual Analyticsin Performance Analysis Kirk Goldsbery Visiting scholar at Harvard, now at ESPN
  • 32.
  • 33.
    Analytics are ReasoningArtifacts … things we use to make decisions. New Data, New Analytics, Same reasoning
  • 34.
    Maps Maps show spatialstructure and patterns Maps provoke spatial reasoning Maps work for all sports
  • 38.
    We’re visual creatures andwhen we see something attractive we want to consume it It takes time to make something that people want to consume. If you were to ask Faulkner how he writes … he doesn't just write, he considers how to frame the story first
  • 39.
    How do youharness spatial science? Sports are spatial Sports are visual Analytics are not spatial or visual
  • 40.
    Spatial Analysis Visualize patternsand quantify information Visual Analytics Translate raw game data into useful information
  • 44.
    Example All LeBron Jamesshots for the last 5 seasons Spatial map of shooting patterns Good for engaging the athlete Good for finding players that are similar or different
  • 47.
    Visualizations can handlebig data As strategic devices As communicative devices
  • 48.
    Two different approachesto visualizations Exploratory Confirmatory Depends on audience
  • 49.
    Sensors lead to quantitativespatial research questions However, provoking spatial reasoning may lead to more questions than it answers Find a question to try to answer and attack it
  • 50.
    Staying Connected: The Risein Fitness Data Questions? Chris Glode GM, MapMyFitness
  • 51.
    Staying Connected: The Risein Fitness Data Chris Glode GM, MapMyFitness
  • 52.
    MapMyRun Team of 100split between Austin & Denver Connected Fitness - social, fun, simple, effective, and rewarding
  • 53.
  • 54.
    160 million workoutslogged in 2013 Team working toward less friction in the app experience
  • 55.
    Growth driven by Smartphones Wireless technology & reduced friction seamless data download Cloud computing Wearables… reduction in hardware costs Obesity epidemic in US
  • 56.
    People use MapMyRunto “outsource their willpower” Friends in the system keep other users more active via notifications
  • 57.
    Techniques for engagement: Games:people keep coming back for competition Games are considered a “jedi mind trick” by MapMyRun, effectively manipulating users to return Route art: motivated users who were otherwise uninterested in social
  • 59.
    Practical applications offitness data Corporate wellness Trainer driven programming, tailored to the individual based on real, recent and new fitness data. Total accountability
  • 60.
    Opportunity via tonsof information mined on the geospatial front Example: advertising to women along certain routes, etc.
  • 61.
    Future Woven wearables Advanced activitydetection Ubiquity of incentives to track fitness iOS7 - Support for passive all day activity tracking in background when app is inactive
  • 62.
    Using GIS tostudy Spatio-Temporal Questions? Patterns in Sport Damien Demaj Geospatial Product Engineer at Esri
  • 63.
    Using GIS tostudy Spatio-Temporal Patterns in Sport Damien Demaj Geospatial Product Engineer at Esri
  • 64.
    Space and timego hand in hand in sport
  • 65.
  • 66.
    Example Who: Federer vMurray Data: 1706 spatial points 3D GIS & streaming video
  • 71.
    The serve: themost important shot Speed & spin: the most important metrics Variation is key: mapping unpredictability is important
  • 72.
    Approach 1. K Meansalgorithm - looks for natural clusters in the data, balls that are close in space but also have a similar attribute 2. Create euclidean lines & calculate large mean distance 3. Tag the most important points in the match 4. Add a feature overlay – Pseudo Realism – putting players back in their environment
  • 75.
    Smart Soccer Qaisar Hassonjee Questions? VPInnovation, adidas Nelson Rodriquez MLS EVP Competition & Game Operations
  • 76.
    Smart Soccer Qaisar Hassonjee VPInnovation, adidas Nelson Rodriquez MLS EVP Competition & Game Operations
  • 77.
    How is technologyenabling sports analytics?
  • 78.
    Wearables Multifunctions, always connected,smart/ aware devices that measure me Enabled by advances in sensor technology, algorithms, data science Everyone from startups to established brands are developing tools What are you looking for? What kind of sensors to develop, ease of use
  • 79.
    Lots of fearin the industry... but lots of copycats once something works
  • 80.
    The adidas TeamSystem Open platform… more sensors over time can be added to the platform Measures Heart rate, Speed, Distance, Location, Acceleration 100 shirts 30 pods 4 ipads
  • 81.
    19 MLS clubsare now using the system - bell curve of adoption
  • 86.
    iPad App Most peopledon’t look at all 20-30 parameters, just the top 2-3 How is the information actionable? How is pre-season trianing improving the fitness of my athletes?
  • 87.
    Results Athletes appreciate andwant to use the technology Tool can extend careers and improve performance Injuries are down 2% this year
  • 88.
    Future Drop the techinto the academy programs & build a national data set of kids & analytics using the system Commercial opportunities, super fan stuff, fantasy teams, etc.
  • 89.
    USA Volleyball Questions? Anton Willert TechnicalCoordinator/Tem Manager US Men’s National Team
  • 90.