Data Analysis: Using data to identify a ball-playing centre back in the J1 League

In this data analysis piece, we will look to identify ball-playing centre backs in the J1 League who may be a target for Ange Postecoglou’s Celtic.

With a league and league cup double in the bag, and a place in the group stages of next season’s Champions League confirmed, Celtic’s full attention will now be turning to this summer’s transfer window.

Although there will not be a huge turnover like last summer, Celtic have several positions that they will look to add to ahead of the next season’s challenges with centre back is believed to be one of those.

Cameron Carter-Vickers and Carl Starfelt, both just recruited last summer, have formed a productive partnership this season, helping Celtic register the lowest goals conceded (20) and expected goals conceded (21.8) in the league.

However, even if Carter-Vickers’ loan deal is made permanent this summer, there is uncertainty regarding the future of third and fourth choice picks Stephen Welsh and Christopher Jullien.

More importantly, there is the question of all four’s suitability to Postecoglou’s system which looks to dominate the ball at a high tempo, taking minimal touches and keeping it moving at pace.

As the Celtic manager put it himself in a mic’d up video in the early weeks, “Ball speed is very important for us, ball speed is number one. Even when we are keeping possession, I don’t want the ball to move slowly. When we pass it, we pass it with purpose.”

Carter-Vickers and Starfelt have both adapted well in general, and both are more than comfortable defensively, but neither can be described as natural ball-playing centre backs though.

Therefore, it is believed that Celtic will look to target a centre back who can be more progressive in this position.

Postecoglou has already had great success recruiting from J1 League, a league he previously won with Yokohama F Marinos. Kyogo Furuhashi, Daizen Maeda and Reo Hatate all played a key role in Celtic’s charge towards the Scottish Premiership title this season.

Yosuke Ideguchi has also been recruited from the J1 League during Postecoglou’s debut season, but injury problems have hampered the early stages of his career in Glasgow’s east end.

With this recent success in this market, the club’s recruitment team will likely still have a close eye on the league for all positions.

Therefore, we will use data to identify centre backs, capable of breaking lines, through their passing or carries, but who are also still defensively sound, in the J1 League who may be a target for Celtic.

Analysis

We’ll start with some basic scatter graphs to narrow down our search. Once we’ve shortlisted a few players we can then use percentile rank pizza charts which will allow us to dig into their data in more depth.

The dataset used for both the initial scatter graphs, and then the pizza charts, include players who have played a minimum of 500 minutes, predominately at centre-back, in this season’s J1 League. All the data is retrieved from Wyscout.

Tableau is used to create the scatter graphs with Python then used to convert the per 90 numbers into percentile ranks for the pizza charts.

Passing ability

First up, we’ll look at a scatter graph for one of the key attributes we are looking for, passing.

General passing stats won’t tell us too much about their profile. We are looking for players who can progress the ball up the pitch through their passing, and, given Celtic’s ball and territory dominance in most domestic games, into the attacking third as well.