How to Get an Analytics Job had Ken Jee on the podcast to talk about being a sports analysts and to give some insight on what can help you land a sports analytics job.
“Getting a sports analytics job is not at all different from getting a traditional analytics job.”
One of the biggest things that I always recommend to everyone is to gather as much experience as you can. Experience does not always have to be from having a role at a company it can be done from volunteering, personal projects among other ways. Especially in sports analytics it’s important to have a portfolio to show that you are interested in the industry and that you have the analytics skills to create value for a team. The projects you create for your portfolio are not only helping a potential employer see your value as an analytics hire but also helping you get stronger and more creative as an analyst. One of the differences between sports analytics versus traditional analytics job is the language and vernacular that is used. Sports analytics has its own language and specifics, including the sports models they use. Rather than always using progression techniques, in sports analytics there are weighted models that people have calculated over time specifically for sports. Take for example the Pythagorean theorem of Baseball, which allows for the projection of win probability (the number of wins that a team might have) in the upcoming year and evaluates how an individual player can contribute to the win total. The best way to familiarize yourself with these sports analytics models, besides just reading about them, is by applying them to projects using available data on Kaggle or other public data sources.
Another key to landing a role as a sports analyst is sharing your work and projects. Whether that sharing is through LinkedIn, twitter, or GitHub. The sports analytics community is super active and therefore, if someone’s work is interesting enough it can get a lot of traction. This traction can often lead to great opportunities and should not be overlooked.
There are three or four different types of roles you can have as a sports analyst according to Ken Jee.
You could potentially work in “the front office” where you’re directly helping teams improve their performance. In these roles you could be advising coaches on their decision-making process where you influence what they’re telling players to work on to improve their game. Then, you also have the other side of working for a sports team where you’re evaluating business factors such as ticket sales. Then you have media and marketing, where it’s less about improving performance and more about telling a really interesting story with the right graphics on TVs during a football game or during a golf tournament. The goal in this role is to engage fans, to bring people in and enhance their viewer experiences by adding to their understanding of the game. Then finally you have the sports betting side with the rise of Daily Fantasy Sports (DFS) along with the relaxation of sports gambling laws. This is a side of sports hat is growing and will continue to grow in the future. This side, according to Ken Jee, is completely about predicting outcomes and not about trying to improve performance. In this role your goal is to understand what happens in a game in the most accurate way to predict what will happen in the next.
The types of models that you will be building depend on what type of role you are hired into. With the first role mentioned, the “front office” type of roles you’re going to look at a lot at player data to try and understand okay what causes positive things or negative things to happen on the field or on the court. If you’re looking from the marketing and media angle, you’re going to build out a lot of exploratory analysis and produce a lot of visuals. Visual influence is huge in the sports world. There’s a reason Instagram is so popular while other not so visual platforms are falling off, because fans and viewers respond best to visuals. For the DFS roles it’s basically all statistical modeling where you’re evaluating outcomes.
Three different types of analytics have been described for different goals within sports analytics at a high-level view. However, looking at the job of a sports analysts on a more day-to-day basis the tasks are not so different from a regular analyst or a data scientist. At the end of the day a sports analyst is still an analyst and that usually involves data collection, data manipulation, a lot of feature engineering, which is how you get the most out of these sports models. A sports analyst is still doing normal model selection, feature tuning, etc. One difference of some roles in sports analytics is that you’re not always communicating information to businesspeople or other data scientists. Sometimes you find yourself in a situation where you are having to communicating to athletes or to or to coaches. While these people are brilliant, they just don’t have as much exposure to the language you use as an analyst. It’s a different skill to communicate information that is complex in a data sense to someone that hasn’t had that much exposure to the field. That is another reason why it can be helpful to use sports terms in a data centered conversation when it involves sports analytics.
There is a portion of data science and analytics where you must be a salesman in the sense that you have to get someone to believe in your analysis. From a business side you usually have to convince a stakeholder to adopt your work to actually use it and put it into practice. As an analyst you need people to see the same things that you’re seeing in the analysis without your domain knowledge. An analyst’s job doesn’t end when the analysis is finished, it ends when that analysis creates value. Accomplishing this comes down to communication and being able to tell a story with the data effectively.