Avatar

Peter Tea

Data Scientist

Sports Media Technology

About me

Hello! My name is Peter and I’m a tennis data scientist at Sports Media Tech. I received my Master’s in Statistics at SFU, where I applied Bayesian models predicting tennis serve direction using tracking data from Roland Garros. This project was supervised by Tim Swartz.

I’ve interned at Aquatic Informatics applying anomaly detection algorithms and at the Canada Revenue Agency studying tax-evasion. I’ve also been an R consultant running programming workshops at SFU Research Commons, and a Statistical Genetics researcher with Kelly Burkett.

You can find my recent tennis analytic blogs and data visualizations on twitter or hosted on-the-t.com:

Hawkeye tracking data is now publicly available! Using shot heatmaps, we dive into player service patterns and find that players overtly target the returner’s backhand on second serve, and that Roger Federer has predictable serve patterns on high pressure points.

How much does height influence a player’s ability to hit aces? Does the added reach also give an advantage on the return game? We model Ace Hit Rates using polynomial regression and also introduce a novel metric Aces allowed rate!

The serve is the most important shot in tennis, and there have been some conspicuous serve performances in recent memory! Eg: Guillermo Coria once averaged 4 double faults per 10 second serves in 2006. In this post, we look at outlying performances and trends in aces/double faults. Interestingly, WTA players on average hit more double faults than they do aces.

In 2019, Alex Zverev’s second serve was a major liability in his game. Would Zverev been better off hitting two flat serves rather than the conventional flat-kick serve combo?

Interests

  • Bayesian Models
  • Tennis
  • Handstands

Education

  • Msc Statistics, 2019 - 2021

    Simon Fraser University

  • Bsc Biostatistics, 2014 - 2019

    University of Ottawa

Projects

Data Science Portfolio

Recent Posts

Serve Power vs. Consistency Tradeoff

Tennis Exploratory Data Analysis

ATP Serve Performances at the Grand Slams

Serving Grand Slams

Visualizing Country Strength in Tennis

Using choropleth maps to visualize nation tennis strength

Presentations

The Analysis of Serve Decisions in Tennis using Bayesian Hierarchical Models

Anticipating an opponent’s serve is a salient skill in tennis - a skill that undoubtedly requires hours of deliberate practice to …

Multilevel Models Applied in Sport

Since the inception of Moneyball, analytics in sport has evolved rapidly - often drawing interest from coaches, broadcasters and fans …

A Genetic Association Study Comparing Kernel-based Methods, with Application to Crohn's Disease

Investigation on Kernel-Based Methods.

An introduction to ANOVA, with application to a Cholesterol dataset

An introduction to ANOVA.

Contact