Examples of previous projects and experience:

---

• For more examples of data engineering projects, see the Data Engineering specific projects in the Portfolio section of this website. Examples of data engineering for football data analysis can also be found in my Football Data Science Portfolio.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• Dashboard to visualise performance metrics of center backs at the 2018 FIFA Men's World Cup using event data from StatsBomb. The code to process this data can be found in the static version of the Jupyter notebook and the corresponding Tableau dashboard can be found here.
• Provided data-driven analysis to clubs and stakeholders as part of Analytics FC's data services including Coach ID and Signature. These analytical pieces and bespoke, static visualisations were created using Python and specifically the matplotlib library for data visualisation.
• For more examples of data visualisation, see the Data Visualisation specific projects in the Portfolio section of this website. Examples of Tableau visualisations can be found in my Tableau Public profile and data visualisation projects using football data can also be found in my Football Data Science Portfolio.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• Example project of an end-to-end machine learning pipeline to train a Chance Quality Model, commonly known as an Expected Goals (xG) mode using Logistic Regression. This model was trained using a historical dataset of 11,000 shots. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, Plotly, scikit-learn, XGBoost, LightGBM, CatBoost, SHAP, HyperOpt, Optuna, and MLFlow). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• Example project of an end-to-end machine learning pipeline to use data to find "the next Gerard Piqué” through Principle Component Analysis and K-Means Clustering. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, scikit-learn). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• For more examples of data science projects, see the Data Science specific projects in the Portfolio section of this website. Examples of data science projects using football data can also be found in my Football Data Science Portfolio.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• Domain knowledge and extensive experience when working with football industry datasets including: event data from StatsBomb, Opta (StatsPerform), Instat, and Wyscout.; tracking data from Second Spectrum, Metrica Sports, Signality; broadcast-tracking data from SkillCorner; and GPS data from STATSports.
• Experienced when working with football analytics libraries including Kloppy for standardising tracking and event data, mplsoccer and ggsoccer for plotting football pitches, socceraction to value the individual actions performed by soccer players, and soccerdata and worldfootballR for scraping football data from sources including Club Elo, ESPN, FBref, FiveThirtyEight, Football-Data.co.uk, SoFIFA, TransferMarkt, and WhoScored.
• Example project of an end-to-end machine learning pipeline to train a Chance Quality Model, commonly known as an Expected Goals (xG) mode using Logistic Regression. This model was trained using a historical dataset of 11,000 shots. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, Plotly, scikit-learn, XGBoost, LightGBM, CatBoost, SHAP, HyperOpt, Optuna, and MLFlow). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• Example project of an end-to-end machine learning pipeline to use data to find "the next Gerard Piqué” through Principle Component Analysis and K-Means Clustering. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, scikit-learn). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• In my free time, I maintain an open source football_analytics GitHub repository of open source analysis, that includes projects for model building including xG models; working with tracking data including the application of pitch control models; the application of action value frameworks including xT, and VAEP; and creating reporting tools using Tableau. This repository also includes a curated list of publicly available resources such as blog posts, libaries, and the latest academic literature, to help people get started with football analytics.
• For more information, see the following article by Training Ground Guru discusses my thoughts and experience regarding football analytics.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• Example project of an end-to-end machine learning pipeline to train a Chance Quality Model, commonly known as an Expected Goals (xG) mode using Logistic Regression. This model was trained using a historical dataset of 11,000 shots. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, Plotly, scikit-learn, XGBoost, LightGBM, CatBoost, SHAP, HyperOpt, Optuna, and MLFlow). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• Example project of an end-to-end machine learning pipeline to use data to find "the next Gerard Piqué” through Principle Component Analysis and K-Means Clustering. This project was completed using Python and key libraries (NumPy, pandas, matplotlib, scikit-learn). The code for this analysis can be found in the Data Analysis and Projects section of my football_analytics GitHub repository and a static version of the Jupyter notebook here as well as in my Football Data Science Portfolio.
• For more examples of football data analysis, see my Football Data Science Portfolio or the Portfolio section of this website.
• See more examples of data analysis of my publicly available data analysis projects created using Python, see the portfolio section of this website.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• See more examples of data analysis of my publicly available data analysis projects created using R and the tidyverse, see the portfolio section of this website.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• Delivered a SQL for Data Analysis Masterclass for Training Ground Guru. This two-hour session provides an introduction to the basics of SQL and an explanation as to what is SQL what is a database; how to create a new database and tables using player performance data; the commands to retrievie records including selecting records, filtering for specific conditions, renaming fields, joining separate datasets based on a unique identifier, ordering data, and grouping and aggregating based on certain conditions; and how to modify existing databases and tables by inserting, modifying and deleting records. To learn more about this seminar, see the Q&A page.

If you would like to find out more:

---

Examples of previous projects and experience:

---

• See more examples of data analysis of my publicly available data analysis projects created using Tableau, see my Tableau Public profile and the portfolio section of this website.

If you would like to find out more:

---

Description

---

A recruitment department of a club are looking to sign a ball-playing center back that fits within their progressive playing style. As part of their recruitment strategy, the department require data-driven identification of the top three ball-playing center backs from the 2018 FIFA World Cup.

Of these three recommendations:

• At least one should be a player aged 23 or younger (at the time of the tournament).
• Due to budgetary restrictions, one player should come from a domestic team playing outside the 'Big 5' European leagues or with a estimated value of less than £5mil as per TransferMarkt.

This project engineers and aggregates World Cup Event data provided by StatsBomb and enriches it with scraped player bio and valuation data from TransferMarkt, to identify the high-performing ball-playing center backs through the recommendation and analysis of suitable KPIs.

From the resulting dataset, professional Football Intelliigence tools are created in Tableau, for use in a professional football club environment, to enable the democratisation of data for football practitioners, helping to reduce their time to insight.

---

The Tableau dashboards can be found on Tableau Public and the Python source code to engineer the data can be found as a Jupyter notebook in my football_analytics GitHub repository. A full slide deck of this analysis along with my other football analytics projects is available in Google Slides.

If you would like to work together on a similar project, email me at edd.webster.analytics@gmail.com.

Description

---

This project takes the publicly available Tracking data for three Hammarby matches in the Swedish Allsvenskan, provided by Signality, to analyse moments in play, create Pitch Control models, and determine the Expected Value-Added by combining Pitch Control models and Expected Possession Value (EPV) frameworks.

This project is conducted entirely using Python in a Jupyter notebook with pandas and matplotlib libraries.

---

The source code for this project can be found in my GitHub repository. A full slide deck of football data science projects, including working with Tracking data, is available in Google Slides. If you would like to work together on a similar project, email me at edd.webster.analytics@gmail.com.

Description

---

This project consists of a selection of Football Intelligence reporting tools, created using publicly available football data from websites and data providers including: StatsBomb, Wyscout, Wyscout, Understat, Signality, and Metrica Sports.

The collection of Football Intelligence tools includes: passing networks, radar plots, xG race charts, shot location visualisations, formation plots, age profile dumbbell charts, and dynamic, dual-axis scatter plots.

The data in this project is scraped, parsed, and engineered using Python in Jupyter notebooks and then visualised as interactive dashboards using Tableau.

---

The dashboards can be found in my Tableau Public profile and the source code to engineer the data used can be found in my GitHub repository. A full slide deck of each of these tools, along with full details as to how they are created and the insight they provide, is available in Google Slides. If you would like to work together on a similar project, email me at edd.webster.analytics@gmail.com.

Description

---

This project takes the publicly available Event data from StatsBomb and Wyscout, to train Expected Goals (xG) models using Logistic Regression and Gradient Boosted Algorithms. From these models, we can then determine the probability of a goal-scoring chance resulting in a goal.

This project is conducted entirely using Python in Jupyter notebooks with pandas for data engineering, matplotlib for visualisation, and scikit-learn and XGBoost for the application of machine learning.

---

The source code for this project can be found in my GitHub repository. A full slide deck of football data science projects, including training Expected Goals models, is available in Google Slides. If you would like to work together on a similar project, email me at edd.webster.analytics@gmail.com.