Data Science with Sam: Sports Analytics in Soccer with Anna D’Souza

概览

This episode introduces sports analytics as a growing area of data science, with a focus on soccer/football. Host Sam frames the conversation around how analytics now appears across sports such as soccer, baseball, cricket, and basketball, and invites Anna D’Souza to discuss her path into soccer analytics.

Anna explains that her route combined playing experience, coaching, economics, Python, football technology, and professional work across women’s football, refereeing technology, game data, scouting, and national-team insights. A recurring point is that analytics in sport is not just technical work; it must be communicated clearly to coaches, players, and other non-technical stakeholders.

The discussion moves from career path to scouting, data privacy, AI, machine learning, computer vision, and advice for newcomers. The main conclusion is that sports analytics works best when data skills are paired with football knowledge, communication, empathy for stakeholders, and genuine passion for learning the game.

分段落总结

[00:03] Introducing sports analytics

[事实] Sam opens the episode of Data Science with Sam and says the topic is sports analytics.

[事实] He notes that analytics has become prevalent across multiple sports, including soccer, baseball, cricket, and basketball.

[事实] Sam introduces Anna D’Souza as a guest who can explain how analytics and data science play a role in soccer.

[推测] The episode is positioned as an introductory but practitioner-oriented discussion rather than a deeply technical tutorial.

[02:07] Anna’s path from player to analyst

[事实] Anna says soccer has been part of her life since early on, and that she played until a foot injury in high school disrupted her playing path.

[事实] After the injury, she chose a more academic route and studied economics.

[事实] She says her economics background introduced her to statistics, Python, data visualization, and econometrics.

[事实] She took an econometrics class using Python and describes that period as difficult because tools such as ChatGPT did not exist.

[03:35] Early analytics work and football technology

[事实] Anna says the 2019 Women’s World Cup was when she began experimenting with soccer data visualization, including shot plots.

[事实] She coached youth soccer during college and obtained a U.S. Soccer C license.

[事实] She later worked in women’s football, then in a referee department where she helped with VAR implementation.

[事实] She says this exposed her to football technology, video/data tools, JSON files, text files, event data, and other unstructured sports data.

[06:57] Professional roles and current work

[事实] Anna says she worked at EA as a data reviewer and later as a data producer, specifically connected to the NWSL side of the game.

[事实] She says she now works on scouting and insights for an agency and also works with the Jamaican Women’s National Team.

[事实] Her current work includes metrics and KPIs aligned with a team’s game model.

[推测] Her career path shows a blend of technical data work, domain expertise, and applied football operations rather than a single linear analytics role.

[07:32] Why experience and communication matter

[事实] Anna describes her career as a roller coaster but says each experience helped her see the game from different angles.

[事实] She says experience is valuable in any field, whether in sport or business intelligence.

[事实] She emphasizes storytelling and communication, saying technical skill alone is not enough if someone cannot explain insights.

[事实] She says coaching children helped her learn how to simplify ideas for both technical and non-technical audiences.

[09:09] Football knowledge as business acumen

[事实] Sam compares sports domain knowledge to business acumen in data science.

[事实] Anna says analysts must communicate with stakeholders such as players and coaches, who may not care about tools like Tableau or Python.

[事实] She says insights should be kept high level and presented in a way that fits the stakeholder’s perspective.

[事实] She argues that having played and coached helps analysts understand game models, coaching sessions, and the support role analysts play inside a team.

[推测] Anna’s view is that sport analytics credibility depends partly on empathy with the football environment, not only on modeling ability.

[12:02] Scouting and data-driven talent identification

[事实] Sam asks how analytics and decision science have affected scouting.

[事实] Anna says national-team scouting variables differ from club scouting variables.

[事实] She says scouting and recruiting are tremendously important, especially for building a program that fits a team’s game model.

[事实] She mentions platforms such as Instat and recommender-style systems that suggest players.

[推测] Data can narrow or structure scouting decisions, but Anna does not describe it as replacing human judgment.

[13:31] Handpicking players and program fit

[事实] Anna says talent identification is important because teams want players who fit what they want to produce on the field.

[事实] She compares player selection to hiring for a job, where the goal is to bring in the candidate most likely to contribute to success.

[事实] She says many institutions have their own talent ID philosophy and programs.

[事实] She contrasts handpicking players through a thoughtful process with an open tryout scenario.

[15:04] National-team scouting and diaspora pools

[事实] Anna says Jamaica depends a lot on its diaspora for player recruitment.

[事实] She says scouting and recruitment are used to find Jamaican players who fit the team’s model.

[事实] She uses the United States as an example of a large country with a central office and regional talent identification structures.

[事实] She says national-team scouting involves identifying what a Jamaican player or U.S. player should represent.

[推测] For national teams, analytics and scouting must also account for eligibility, geography, and federation identity, not only performance metrics.

[16:54] Sports data privacy and compliance

[事实] Sam asks whether sports analytics has regulations, compliance, or data security rules similar to finance.

[事实] Anna says sports data includes event data, tracking data, and wearables.

[事实] She says Europe has data privacy protection rules that are important to follow, especially for wearable data involving human beings.

[事实] She says in the United States the situation can depend on the state, and she mentions California as more similar to Europe.

[推测] The privacy risk is higher when the data is tied directly to athlete bodies, movement, or personal performance.

[18:27] Ownership, data release, and security

[事实] Anna says data release and distribution require caution, especially in leagues such as MLS.

[事实] She mentions collective bargaining agreements as a possible place where data ownership and sharing rules may be defined.

[事实] She says cybersecurity also depends on the hosting provider or platform, such as AWS or Oracle.

[事实] She emphasizes that analysts are dealing with human beings, so the distinction matters.

[推测] The discussion suggests that sports organizations need both technical security practices and clear governance around who can use athlete data.

[20:22] AI and machine learning in sports

[事实] Sam asks about AI, machine learning, LLMs, and computer vision in sports analytics.

[事实] Anna says AI and machine learning are already implemented in sports.

[事实] She mentions analytics companies that acquire event data and betting companies as examples of organizations already using these techniques.

[事实] She says computer vision can track off-ball activity, while on-ball data may involve tagging and synchronization with tracking data.

[22:19] Event-data synchronization and predictive models

[事实] Anna says machine learning techniques can help synchronize event data and tracking data.

[事实] She says StatsBomb and FIFA have used this kind of approach, and mentions FIFA’s EFI program in Qatar.

[事实] She says betting models can be trained with variables such as goals, goals against, goal difference, coach, and opponent.

[事实] She mentions machine learning tools or approaches such as XGBoost and scikit-learn in the context of prediction.

[23:40] Graph neural networks and future research

[事实] Anna says graph neural network research has explored predicting counterattacks and goals.

[事实] She says she is not the most experienced person in graph neural networks or computer vision, though she can speak about them generally.

[事实] She says predictive analytics is more familiar to her.

[推测] The most advanced AI topics in soccer analytics are presented as active or emerging areas, while traditional predictive analytics is more established for her work.

[24:22] Penalty prediction example

[事实] Sam asks about using computer vision or analysis to help a goalkeeper anticipate penalty shots.

[事实] Anna says someone did similar work and names Tyler Heaps.

[事实] She says the work involved predicting probabilities of where a player would shoot a penalty, possibly using a heat map.

[事实] She connects that work to the 2019 Women’s World Cup.

[推测] This example illustrates how analytics can move from post-match reporting into direct match preparation.

[26:05] Advice for newcomers

[事实] Sam asks Anna for advice for people who want to enter sports analytics, including people without coaching or playing experience.

[事实] Anna says the best way is to be passionate.

[事实] She quotes Pelé on success requiring hard work, perseverance, learning, studying, sacrifice, and love for what one does.

[事实] Her advice is to be dedicated, studious, love the game, and love to learn.

[推测] For Anna, curiosity and sustained effort matter more than having a perfectly traditional sports background.

[28:00] Closing and future discussion

[事实] Sam thanks Anna for sharing her journey from economics into sports analytics.

[事实] He suggests they may continue the discussion in the future, possibly with more sports analytics voices in a panel format.

[事实] He tells listeners the podcast is available on YouTube, Amazon Music, and Apple Podcasts.

[事实] He closes by asking viewers and listeners to follow and subscribe to Data Science with Sam.

播客点评/总结

[推测] This episode is valuable for listeners who want a practical overview of soccer analytics careers, especially those coming from data science, economics, or another quantitative field. Its strongest point is Anna’s emphasis on the combination of technical skill, football context, and communication.

[事实] The conversation covers a broad range of topics: career development, coaching knowledge, scouting, national-team recruitment, data privacy, AI, computer vision, betting models, and advice for beginners.

[推测] The limitation is that many technical topics are introduced at a high level rather than developed in detail. Listeners looking for model architectures, data schemas, or implementation walkthroughs will likely need additional resources.

[推测] The episode is best suited for early-career analysts, data scientists curious about sport, and soccer practitioners who want to understand how analytics fits into team, scouting, and performance environments.