Eric A. Eager - Football Analytics with Python & R. Early Release

With Early Release ebooks, you get books in their earliest formthe authors raw and unedited content as they writeso you can take advantage of these technologies long before the official release of these titles.

This will be the 1st chapter of the final book. Please note that the GitHub repo will be made active later on.

If you have comments about how we might improve the content and/or examples in this book, or if you notice missing material within this chapter, please reach out to the editor at .

Football analytics, and more broadly, data science, require a broad setof tools. Successful practitioners in these fields require anunderstanding of these tools. Statistical programming languages are abackbone of our data science toolbox. These programs allow us to cleanour datasets, conduct our analyses, and readily reuse our methods.Although many people commonly use spreadsheets (such as Microsoft Excelor Google Sheets) for data cleaning and analysis, we find spreadsheetsdo not scale well. For example, when one has to work with large datasetslike tracking data, which can contain thousands of rows of data perplay, spreadsheets simply are not up to the task. Programming languagesalso allow for easy reuse because copy and pasting formulas inspreadsheets can be tedious and error prone. Lastly, spreadsheets allowundocumented errors. For example, spreadsheets do not having a method tocatch a copying and pasting mistake.. Furthermore, modern data sciencetools allow code, data, and results to be blended together ineasy-to-use interfaces. Common languages include Python, R, Julia,Matlab, and SAS. Additional languages continue to appear as computerscience continues to advance.

As practitioners of data science, we use R and Python daily for ourwork, which has collectively spanned the space of applied mathematics,applied statistics, theoretical ecology and, of course, footballanalytics. Of the languages listed previously, Python and R offer thebenefit of larger user bases (and hence likely contain the tools andmodels we need). Both R and Python (as well as Julia) are open source.Open source means two types of freedom. First, anybody can access allthe code in the language, and this freedom is sometimes called librefreedom (think free like in free speech). This allows volunteers tohelp improve the languages, such ensuring that users can debug the codeand extend the languages through add-on packages. Open source alsooffers the benefit of free to use for users, sometimes called gratisfreedom (think free like in free beer). Hence, users do not need topay thousands of dollars annually in licensing fees. We were initiallytrained in R, but have learned Python over the course of our jobs.Either language is well suited for football analytics (and sportsanalytics in general).

We encourage you to pick one language for the book and learn thatlanguage well. Should you need to learn a second programming language,it is easier if you understand the programming concepts behind a firstlanguage well. Then, you can relate the concepts back to yourunderstanding of your original computer language. Although many peoplepick favorite languages and sometimes have arguments with each otherover which coding language is better (similar to Coke versus Pepsi orFord versus General Motors), we have seen both R and Python used inproduction and also used with large data and complex models. Forexample, we have used R with 100 GB files on servers with sufficientmemory. Both of us began our careers coding almost exclusively in R, buthave learned to use Python when the situation has called for it.