Sports analytics I

class: center, middle, inverse, title-slide

# Sports analytics I
### GEOG 30793
### May 28, 2019

---

## Today's agenda

* Data analytics: an overview

* Sports analytics: an emerging field

* Case study: sabermetrics and baseball

* Probability and statistics

* Case study: probability and football

---
class: middle, center, inverse

## Data analytics

---

## What is data?

<img src="http://40.media.tumblr.com/3d0477bad4077debabcdbf29a2df3670/tumblr_n6cjvqfzgr1td9006o1_1280.jpg" style="width: 850px"><figcaption>Source: [bigdatapix.tumblr.com](http://bigdatapix.tumblr.com/)</figcaption>

---

## The data analysis process

<img src="http://r4ds.had.co.nz/diagrams/data-science.png" style="width: 800px"><figcaption>Source: Wickham and Grolemund, [_R for Data Science_](http://r4ds.had.co.nz/introduction.html)</figcaption>

---

## Data science

Emerging (and highly-paid) field that includes:

* Data exploration, preparation, and visualization

* Statistical modeling and machine learning

* Software development and engineering

* Practical business and/or domain expertise

---
class: middle, center, inverse

## Sports analytics

---
class: middle, center, inverse

## Discussion: how might you use data in the sports domain?

---

## History of sports statistics

.footnote[Source: [onsports.wordpress.com](https://onsports.wordpress.com/2007/10/10/agate-is-essential-to-covering-sports/)]

---

## Analytics in sport

.footnote[Source: [SportTechie.com](https://www.sporttechie.com/fifa-will-provide-teams-with-in-match-tracking-video-during-world-cup-soccer-football/)]

---

## Analytics in sport

.footnote[Source: [JumpForward](http://www.jumpforward.com/predictive-analytics-college-athletic-departments/)]

---
class: middle, center, inverse

## Case study: Sabermetrics and baseball

---

## Sabermetrics

---

## Bill James

.footnote[Source: [worldseriesdreaming.com](http://worldseriesdreaming.com/tag/bill-james/)]

---

## Wins above replacement (WAR)

* Composite metric that measures __the number of additional wins a player is worth__ relative to an average player at that position

* [2019 WAR data: ESPN.com](http://www.espn.com/mlb/war/leaders)

---

## Weighted on-base average (wOBA)

* Batting average: how frequently a batter gets a hit

* On-base percentage (OBP): how often a batter reaches base

* wOBA: different weights applied to different on-base appearances based on _value_

---

## Defense-independent pitching statistics (DIPS)

* __Earned run average__ (ERA): earned runs per nine innings pitched

* __DIPS__: focus on statistics _specific to the pitcher_

* Example: fielding-independent pitching (FIP)

---

## The landscape of baseball analytics

Theo Epstein ([source](http://www.chicagotribune.com/sports/baseball/cubs/ct-spt-cubs-theo-epstein-analytics-20180327-story.html)):

> There's no doubt the landscape is a lot flatter now... Most organizations are operating basically with the same data streams, same numbers.

> So if everyone has the same information, you want to put a premium on a humanistic approach understanding people... Putting them in a position to succeed and supporting them as human beings and individuals and the chemistry of the group as a whole.

---
class: middle, center, inverse

## Discussion: the value of data and analytics in baseball

---
class: middle, center, inverse

## Probability and statistics

---

## Probability

* **Probability**: the likelihood of the occurrence of an event

* **Statistical significance**: the probability that an observation/effect is not the result of random chance
---

## Frequentist statistics

* What is the mathematical relationship between an outcome variable `$Y$` and one or more other "predictor" variables `$X_{1}...X_{n}$`?

The linear model:

$$ Y = Xb + e $$

where `$Y$` represents the outcome variable, `$X$` is a matrix of predictors, `$b$` represents the "parameters", and `$e$` represents the errors, or "residuals"

---

## The normal distribution

---

## Bayesian statistics

* Given our _prior beliefs_ about a mathematical relationship, what is the probability of an event occurring?

.footnote[Source: [AnalyticsVidhya.com](https://www.analyticsvidhya.com/blog/2016/06/bayesian-statistics-beginners-simple-english/)]

---

## Machine learning and predictive modeling

* The use of statistical modeling to _make predictions_ about future behavior

Example: [Amazon recommendations](https://www.theonion.com/amazon-com-recommendations-understand-area-woman-better-1819568885)

---

## Probability and football

.footnote[Source: [FiveThirtyEight](https://fivethirtyeight.com/features/kickers-are-forever/)]

---

## Probability and football

* Example: [New York Times 4th Down Bot](https://www.nytimes.com/interactive/projects/4thdownbot/team/cowboys)

---

## Probability and football

.footnote[Source: [FiveThirtyEight](https://fivethirtyeight.com/features/kickers-are-forever/)]

---

## "The team that never punts"

---

## The 2017-2018 Eagles

.footnote[Source: [deadline.com](http://deadline.com/2018/02/super-bowl-ratings-eagles-patriots-this-is-us-nbc-1202278181/)]

---

## The Eagles and 4th-down conversions

---

## 4th-down conversions in the NFL

* [Statistics: ESPN.com](http://www.espn.com/nfl/statistics/team/_/stat/downs)

---

class: middle, center, inverse

## Next up: Sports analytics II

<style>

body {
  font-family: Verdana; 
}

h1, h2, h3 {
  color: #840000; 
  font-family: Verdana;
  font-weight: bold; 
}

a {
  color: #ff0000; 
}

.inverse {
  background-color: #840000;

}
</style>