My name is Michael Morrissey, an Irish born experimental physics and data scientist. All my life I have been interested in how things work, why do they work, what are the physical processes that make it work. This led me to a career in physics where I received a PhD in experimental quantum optics. This allowed me to work while also doing a little exploration of the world where I worked in Germany, Greece, South Africa and Czech Republic. In early 2019 I changed career from being a research physicist to being a data scientist. Data science uses scientific methods and processes, algorithms and systems to extract knowledge and insight from data. It employs techniques and theories drawn from many fields within the context of mathematics, statistics, computer science and information science. Given my new found passion for this field and newly developed skills, I became interested in combining this with an already existing passion… SQUASH. However, after thoroughly scouring the internet & research literature, I realized that there is only a limited amount of work performed in squash data analysis that is publicly available compared to other major sports who invest a lot of money in maximizing player performance through sports science and data analysis. If you look at tennis there is an overwhelming amount of statistical analysis performed on all aspects of the game, players, court types, shot selections, shot techniques etc etc. Of course tennis is much more lucrative than squash and so it makes sense that more effort is made to give athletes a better opportunity to succeed. The next step of the Squash Stats journey was to determine if there was a sufficient amount of publicly available data in order to perform a reasonable amount of statistical analysis. Fortunately with sites like Squash Info and PSA World Tour, which provide ample years of data such as match results, rankings, prize money, player profiles, sponsors, there is an abundance of data available. Therefore, I decided to start a project which brought sports analysis to the world of squash – and this resulted in the birth Squash Stats.
Off to a Good Start
The first month or so was spent gathering data from the internet, constructing a database, building a software infrastructure and performing initial data analysis. I decided to post my findings on Facebook with a simple post with graphics to display the results of my analysis with a small description. Slowly Squash Stats started to gain followers who were happy to see content on a sport they love and the Facebook-type forum allowed them to add comment, provide feedback, suggest new ideas for analysis and share content which is essential to the growth of Squash Stats.
Growing the Team
In September of 2019, just six months after its inception, the Squash Stats team doubled in size when Eric Katerman joined the project. With Eric’s experience in data engineering, mathematics and programming, he converted the software aspect of Squash Stats into a much more functional, structured and professional system, which is essential to the maintenance and development of the Squash Stats project. Eric’s data science interests are primarily in machine learning and computer vision, and he is currently developing tools to extract data and perform analysis on squash video. Thanks to recent developments in open-source AI and computer vision software, squash video analysis tasks are less daunting now than they were just a couple years ago, and this enables researchers like ourselves to build systems that extract fascinating, valuable insights from video footage of squash circuits, drills, and match play.
The Long Term Vision
Squash Stats will continue to perform statistical analysis on matches, games and tournaments as it is currently doing. In addition, we hope to push the boundaries of what analytical tools are available to analyze squash statistics and data. This includes following advances to AI and applying cutting-edge computer vision techniques and machine learning algorithms to publicly available squash data. We are also interested in putting real statistics behind answers to questions like “who plays tiebreakers best?” and “how often does a player win after having a two-point cushion at 8-6?” and “how important is it to get to five points first in the fifth game?”. Clearly much great work has been done on analyzing squash; however, we feel that squash still lags behind other sports such as baseball and tennis that have an overwhelming abundance of statistics available for public consumption. At the end of the day, we are passionate about the sport of squash and data science, and we are excited to work towards our mission of Advancing Squash Through Statistics.