Some Personal Reflections on Getting Started in Soccer Analytics

Yesterday Ravi (@Scribblr_42) wrote a great microblog titled “Is Fanalytics Intimidating?” and you can find it here:

It’s really interesting, and you all should read it and then come back to my post. I’ll wait.

This isn’t a methods post about how to get started, but I wanted to share some of my experiences on breaking into the analytics community and maybe offer some advice to people who want to get involved.

I’ve been doing this about a year now, and converted my personal Twitter account into a “soccer only” account around 7 months ago. In that time I’ve grown my following from about 110 friends and former students to over 2500 followers, and feel incredibly fortunate to have picked up a good-sized following in such a short time.

I’ve posted it before, but this started as an excuse to improve my data science skills that I didn’t think would go anywhere. I likely wouldn’t have kept up with it if there wasn’t an audience for what I’ve been doing – I’ve briefly tried political blogs in the past but they never really caught on and I don’t feel like I have anything unique to add to the political blogosphere.1 I don’t see the purpose in saying something that’s already been said for a couple dozen people to read, so I’ve never really stayed with it. But with soccer I’ve built up an audience for my work so I’ve continued to put out content that hopefully people continue to enjoy.

The biggest thing that has helped me has been bigger accounts sharing my work, and I couldn’t be more appreciative to the people who have done so. Mike Goodman has been particularly supportive from the beginning(@The_M_L_G), as have @GoalImpact, and @7amkickoff. I’m incredibly grateful to the people who regularly share my work, particularly Jake Kilov (@Kilonater3000) and Naveen Maliakkal (@njm1211) who have consistently retweeted me for a while, and anyone else who does.

Since I moved in to Women’s Soccer, @DasGherkin has been incredibly generous in promoting my account, and I’ve picked up over 200 followers just from her recommendations. She’s kind of a big deal in the WoSo Twitter world, and for her to share my work has given me a real credibility boost  in that world and has been invaluable in helping me get the word out for my upcoming NWSL analytics work. I know I’m leaving a bunch of people out, but I’m grateful to everyone who has helped me out.

But it’s not just about growing my audience – simple support from people in the community has meant a lot and kept me going. Tom Worville (@Worville)  helped me with coding and data issues more than a few times, especially in the beginning, and James Curley (@jalapic) and @UTVilla have both given me great help and have helped my R programming skills grow exponentially. And even something as simple as seeing prominent accounts favoriting my tweets has given me the motivation to go forward and keep doing work. Favoriting is especially important because it’s a costless act beyond pressing a button, but it’s a nice validation that I’m doing something interesting/good and a clear sign that someone’s reading it. The community embraced my work early, so it has been less intimidating for me than it has been for others, but I can see that it would be intimidating if someone doesn’t get this support. I don’t know if I would still be posting public analytics work without it, and I’m someone who was already confident in my analytics skills.

To people with a lot of followers: take the time to engage with people who don’t have as many eyes seeing their work. You don’t have to have long conversations (although it’s nice), but something as simple as a favorite, or even a retweet, can mean a lot to someone trying to find a place in the community. I do this every so often with my Retweet Days where I share work from people who have fewer followers than I do. And the best part is that the more followers I get, the more voices I can amplify. I’m happy to do it.

I’ve also been trying to give back in terms of my “Intro to Analytics” YouTube Course. People seem to be enjoying it, and I need to add some more chapters as soon as I can find the time and energy to do so. Maybe people who watch these will become involved in analytics, or maybe they’ll just be more able to participate in the conversation, but hopefully this will let more people get involved and make people a little less hesitant to participate. At some point we all started out with virtually no followers, so why not pay it forward and try to help people who might be part of the next generation of soccer analysts?

Final thoughts: for those of you looking to get involved,  don’t worry about the math. Learn some simple concepts from my YouTube channel (if you don’t know them already), find some public data, and go from there. I know my work focuses on predictions, but like someone said on Twitter I don’t think there’s much more room in that space unless you can out-predict my model and the other ones that are out there (which will be tough). Same goes for Expected Goals: I don’t think there’s a lot of room for new xG models in the analytics world unless you can put something together that significantly beats the prominent ones already out there. But there’s lots of work to be done, so look for those gaps and fill them.

Most importantly, be a good citizen, regardless of how many followers you have. If you’re doing interesting work, engage with other people’s work in a positive way, and write often you’ll have a good chance at building a following. Even if you don’t, you’ll have some positive experiences and share your work with like-minded people. Hopefully people wanting to get started can take my advice, and hopefully people who have a strong following can help encourage new people to participate in the community. It really helped me get a foothold and feel like my work was being appreciated, and the teacher in me wants to help others along the same lines.


  1. I do original research as part of my job, but that’s academic, not blogging.

Leave a Reply

Your email address will not be published. Required fields are marked *