On Data & Football

Motivations and Brief Project outline.

Featured image

Football for me is an expression of a craft that is one of the only true meritocracies in society. Given that world view, I have always had a deep appreciation of the players who made the sport look effortless.

Growing up, one of the players I did my best to try and emulate on the school pitches and playground was Andrea Pirlo. I don’t really believe in having heroes but this man was as close to an idol for me, more than any other sportsperson. I would religiously study the way he played on midweek champions league nights and often try (and fail) to replicate the passes and body-feints on a Saturday morning.

pirlo

One of the things that really drew me to this player was his change of position. At face value, this is normally taken as an isolated case of an extremely talented footballer who had many strings in his bow, however when you peel the curtain back, many household names have achieved similar feats when changing position. It got me thinking, how many more players could have had significantly more impactful careers but were wasted in a position that did not align to their strengths, how much money could clubs have saved in years gone by by moving a player around, instead of completing an expensive yet very risky transfer. Here are a few household names that started out in other positions that may surprise you:

Players

I made this blog to effectively answer this question of :

‘Can we identify a more suitable on field position for any given (outfield) player using machine learning’.

The excitement of combining two of my favourite subjects had me running away with so many ideas, some more practical and realistic than others, but honestly with very little experience in the combined field itself, I believed it was best to just start from zero and build my way up in terms of understanding and document the entire process. So I promised myself to put some time into really getting my teeth into this before life takes the front seat. I have an actual job to maintain as well as friends and family. So it took me a bit of time to figure out how best to explore this but I eventually started to create the blog to learn in public and dive right in.

As I see it, the road map I envisage for this blog will start with; acquiring the right, high quality data. I will be documenting what data sources I’m using and why. Next, getting to grips with data analysis & data visualizations will also heavily feature in the subsequent tasks I will be blogging about. Hypothesis testing will also be very prevalent in my posts; is this even a valid argument to even explore & finally how can I apply modern techniques of machine learning to answer the eventual questions that will arise from this learning journey.

Most importantly, I want to note that this blog will include my learnings and my take-aways so far. I am no expert in programming nor a professional football analyst, so take what I say with a grain of salt and double-check with the experts and please correct me where you know that I am wrong or things you think I’ve missed off.

With all that said and done, let’s get building!