Tableau: How much data are you collecting and what were you doing with it?
Brian: We've been taking unstructured and structured data in from various sources for a long time now. The problem has been most of it has been unusable. So we've had this massive amount of data just collected, and we've not been able to do anything with it.
We're collecting over 6,000 XML files every single season, and to make any sense of that, to actually make decisions from a player recruitment perspective, we needed the power of Alteryx, and the user-friendliness of it as well, to be able to actually parse the data. And then not just parse it but actually use predictive analytics tools to actually gain deep insight and to actually understand which players are going to benefit our club.
Tableau: What kind of support have you gotten in using the software?
Brian: The help we've had, the hands-on help, I suppose, we've had from the Information Lab, The support we've had from both Tableau and Alteryx has been fantastic, both across the phone, e-mail, and in person. It's been great because it's enabled us to create not just the basic-level dashboards that perhaps we would have done if we didn't have that help, but we've been able to advance our work and adapt it to make it very relevant to the individuals and to be able to deliver the insights that we actually want as a club.
Tableau: George, you had a few questions for Brian?
Alteryx president George Mathew: So what are some of the things that you're getting from that XML data itself? What's the source content look like these days, particularly for the sports analytics that you're starting to drive?
Brian: We get location information. So that's based upon, you know, where are events happening, what the type of event is, whoever is challenged or unchallenged, who's involved with it, and the outcome of that.
Now, that, to us, provides a very basic level of information, but we need to delve deeper, look at the relationships, look at where on the pitch things are actually happening, so not just an X/Y coordinate but actually is that X/Y coordinate within a penalty area. Is it within a six yard box? Is it within a zone that's actually a good goal score in efficiency zones? That's a benefit that the special analytic tool has allowed us within Alteryx.
George: You mentioned that the pitches are different in every situation. How do you normalize that? How do you make it seamless to blend that information as you're then doing the visual analysis in Tableau?
Brian: Every single pitch in football is different. So when we get these XML files, and you can't just say, “Right, if the ball is here, that means it's in the penalty area, if the ball is there it's in the sort of on the halfway line,” because everything is different.
Tableau: How do you turn those basic coordinates into something more relevant?
Brian: We have to use the sort of some generic coordinates, which show the sides of the pitch, so we can then draw a pitch using the tools within Alteryx. Not only then draw the pitch but draw the halfway line, draw the 18 yard box, draw the six-yard box and any of the zone that we may want to. Like, we split the pitch into left, central and right, as well as into thirds, going up the pitch. So we're able to do that using the spatial tools, and then use spatial match just to identify where events are happening within those areas.
Tableau: How did you manage all this information before you adopted Tableau and Alteryx? Where you able to digitally recreate the pitch?
Brian: Prior to having the XML is we'd get CSV file, and that basically provides us with very generic information about what's going on on the pitch.
The key thing it doesn't provide is locations, locations, X/Y coordinates. We don't get that, which means that realistically we don't know how effective a pass was, how effective a cross was. Whereas actually having that location, having those coordinates and making sense of them in Alteryx means that we can start to create strategies for training, right? Have the team done what we were asking of them? Is the pattern of play what we wanted? When we're looking at opponents, what's their pattern of play, how can we counteract that?
And it's all actionable insight, it's all things that actually influence training on a daily basis and influence not only team selection, but team strategy for match day.