Collaborations

Soccermetrics (i.e., Howard Hamilton, the founder) is seeking to establish collaborative research projects on soccer analytics with interested parties.  The goal of the collaborations is to perform fundamental research on a wide range of problems in soccer analytics and related analytical problems in other sports that is of publishable quality in peer-reviewed journals and compelling to practitioners in the sports industry.  By working with subject matter experts in other fields, progress can be made on these projects more efficiently and all members can learn from each other’s expertise.

This page presents an overview of the research collaboration structure with answers to the major questions.  Collaborations for specific years will be described on separate pages.

Who are you looking for?

Basically, I’m looking for collaborators who are interested in one of my proposed projects (or, have one of their own) and have the complimentary expertise to contribute to the project.  I expect collaborators to contribute significantly to the analysis part of the project, so your math and statistics skills need to be very sharp.  It’s likely that most collaborators will be academic researchers, but there might be amateurs and autodidacts who fit the bill.  I’m especially interested in working with collaborators from backgrounds that aren’t well-represented in STEM in general or sports analytics in particular.  Regardless, the best collaborations will be with partners who have the technical skills and the desire to work hard and produce rigorous and high-quality output.

What will a collaboration look like?

The details of these collaborations are being fleshed out, but regular and open communication will be fundamental to a successful effort.  Communication will take place via email, phone calls, or text/video messaging.  Data and software code will be shared and worked on using GitHub.  Face-to-face travel is more expensive and may not be workable depending on where collaborators are, unless we live close enough together or happen to be at the same conference.

How long will a collaboration last?

A research collaboration will be expected to last 12-18 months.  This time covers the amount of time required to collect, clean, and organize data (which will always take more time than expected), perform iterative analyses or experiments, and write up, present, and publish results. In order to ensure that all collaborations can be performed at a high level of quality, no more than two collaborative projects will be performed at the same time.

What is the desired end result?

The goal of these collaborations is to produce results of publishable quality in peer-reviewed journals.  Candidate journals include the Journal of Quantitative Analysis in Sports, Journal of Sports Analytics, or other relevant publications.

Along the way, results will be communicated to the sports analytics community via blog posts (on Soccermetrics and/or elsewhere), social media channels.  It is expected that in-progress and final results will be submitted to sports analytics or other subject-matter conferences of comparable quality to the MIT Sloan Sports Analytics Conference (SSAC) or New England Symposium on Statistics in Sports (NESSIS).

It is desired that all software created in the course of these collaborations be released to open source, and that the data used in the research be made publicly available.  (An exception will be made for data that is proprietary to the data supplier.)

How do I express interest?

Send an email to [email protected] with the phrase “Research Collaboration” in the subject line.  State which collaborative project is of interest, or present a detailed proposal of your own.   Attach a copy of your CV and/or portfolio of previous analytics projects.