Visualising Actual Performance Compared to Predicted Performance

Sydney_moderns_cover_72dpiRGB.jpg.300x1200_q85Introduction

The Sydney Moderns Exhibition in general, and Roy de Maistre in particular, have sent me off thinking about visualisation this week.

I have been wondering if I can combine my interest in the impact of a team’s ranking on performance with a narrative form that has a visual impact.

Thanks to an alert from my son, Sam, my colour search was transformed by the ColourLovers site. I was able to use their palette tool to explore HTML Colour Codes.

I decided to create two single hue palettes, one for winning (blue) and one for losing (red) performance. The ColourLovers palette enables the user to set the hue (H value), its intensity (saturation) (S value) and the darkness of the colour (V value).

I chose a blue hue for winning performance. The information about the hue includes: the Hex Code; the Red (R), Green (G) and Blue (B) values and the HSV values:

Win Code

and a red hue for losing:

Lose Palette

I have used the performance of two AFL teams in the regular 2013 season to explore the impact colour might have on visualising performance.

Hawthorn and Richmond

Hawthorn were the minor premiers in 2012 and repeated this performance in the 2013 season.

With their number 1 ranking for the 2013 season, this was the prediction of their performance (a completely blue profile):

HPredict

The darker the blue (indicated by higher saturation values) the closer Hawthorn’s opponents are to them in ranking.

Hawthorn’s actual performance was:

H Actual

The darker the red, the closer the team defeating Hawthorn was in ranking terms. The loss to Richmond was a very interesting result. Richmond were a significantly lower ranked team but as the visualisation of their season below indicates, they were very competitive in 2013.

A comparison of Hawthorn’s predicted and actual season underscores how comprehensive their minor premiership win was.

H Compare

Richmond were ranked 12th in the 2012 regular season. The prediction for 2013 based on this ranking was:

R Predict

The darker blue and red hues indicate a close proximity to teams in terms of ranking. This prediction suggested two wins in the first eight games with an opening game against Carlton ranked two positions above Richmond.

Richmond’s actual performance was five wins starting with a first round defeat of Carlton:

R Actual

The comparison of the predicted and actual indicates how much Richmond improved this season:

R Compare

This comparison prompts me to think about Richmond’s progress by defeating teams ranked near to them.  These colours have high saturation figures and appear as darker colours in the graphic. The highlight is the light blue colour of the win against Hawthorn in Round 19.

Discussion

This is a first attempt to explore my use of colour in visualsation. I am hopeful that a narrative is present in these visualisations. I have made a conscious decision to use a single hue for winning (blue) and for losing (red). I think the Hawthorn images indicate how successful their season was. After a Round 1 defeat, they followed their predicted path (based on ranking) to Round 15.

I hope that the Richmond visualisations indicate a change in their fortunes in 2013. Their season was a much bluer season that predicted. They were able to defeat closely ranked rivals. Their victory against Hawthorn in Round 19 was one of the performances of the season for me.

I look forward to developing this use of colour whilst being mindful of the excellent advice available about what constitutes better practice.

Photo Credit

Sydney Moderns Catalogue Cover (Art Gallery of NSW)