Junking football

As a geographer, professional researcher and the son of a graphic designer* I’ve long been interested in the field of data visualisation, which drew me to Kaiser Fung’s excellent Junk Charts blog. It’s made me about the way that I report data to a non-specialist audience and reflect on whether my work passes the junk chart trifecta checkup. I was agog when I saw this chart last week, and felt compelled to write this analysis in Junk Chart style.

I was on a business trip to Dubai last week, where my heart skipped a beat at breakfast when taking in the magnificence of this ‘big graphic’ in The National newspaper. (The original Graphic News chart is hard to read on-line, so the scan below will have to do.) It seems to fail all three parts of the Junk Charts Trifecta Checkup: it’s not clear what question is being answered; there’s inconsistent presentation of one piece of data and another key data point is missing which therefore obscures a key piece of the narrative; and it’s near-impossible to establish what the chart is saying.

As a researcher my passion lies in data analysis rather than football (soccer), but improving this chart was an open goal, so I worked up an alternative interpretation.

I suspect that the (pie? radar? bullseye?) chart is trying to show that a high proportion of English Premier League football club revenues are devoted to paying players’ wages, and that most of the clubs are operating at a loss (there is no salary cap in European football). However it is unclear what the light and dark colours for each segment indicate, even when reading the key hidden on the left, and the use of colour – based very loosely on the team kits – further confuses the reader.

In addition, profit/loss is provided for only seven of the twenty clubs, mostly away from the chart itself. I have no idea why Blackpool sticks out, perhaps the area-based segmentation means that its slither was too thin to read unless it was magnified.

The most interesting piece of data is missing from the chart: how do income, wages and profit/loss affect performance? Last Sunday was the finale of the league, so I went back to the data source and final league table to see what the data actually tell us. The reworked chart on the right adopts a minimalist black and white style, and looks better for it. It is now easier to spot the gap between wages and turnover, profit/loss data for all teams are provided**, and the impact of all three figures on league performance is apparent. (Sixty data points for twenty teams are difficult to fit onto one chart, the vertical layout is a compromise; I toyed with adding the number of points given the skewed distribution, but this over-complicated matters.)

The takeouts from the chart are that club owners need very deep pockets to reach the top of the table; that outside the top six clubs there is no correlation between wage bill and league performance; and that the top clubs will have a lot of trouble complying with new financial regulations. Oh, and sometimes a graphic can be more attractive if it chooses to dress down.

I’m sure there’s room for improvement, but I’m pretty sure the black and white version is a step forwards.

*This makes me sound like an arrogant prig. I’m not!

**I chose to tweak the figures for Arsenal as their profit was distorted by a very large one-off property transaction.This was removed from their income, and more recent profit/loss data used. It’s a bodge, but probably a necessary one. Ideally it would show financial and league performance from the same year, the attached is a compromise last year’s financial data and this year’s performance, required as I don’t have access to financial data for the teams that were relegated in 2010.