The dataset provided by the client includes several parameters of netball games in the UK. It comprises the ranking of each team along with the scores, and the number of passes from three positions- centre, interceptions, and restarts. Although it helps to understand the factors of victory, we strongly believe additional data, along with useful models, is required for more in-depth analysis. 

​

I/ RECOMMENDED MODELS

We recommend models and visualisations to be developed, in order to make the most of the available dataset. Although each of these models has a different purpose, they all intend to set priorities for the Sirens and to adjust the training of team members.

​

1. A board comparing statistics for Centre/Restart/Live and for Passes/Shots/Goals between Sirens and the average of the first 5 teams. We think that this comparison with the leading teams is relevant because it sets the aim of the Sirens. This overview will show clearly where Sirens must improve.

2. A process chart describing the efficiency of the Sirens in terms of passes, shots, and goals as a proportion of the overall score of the game. These visualisations reveal the game phases the team should spend more time practising on. For instance, if the ratio shots overpasses seems to be critical in the final ranking, Sirens may want to practise further ball possession.

3. A team profile not only for the Sirens but also for each team of the league. This “card” provides the main performance indicator of the team such as its victory rate, accuracy rate, and proportion of each passes. This model would help Sirens to compare themselves to the best teams.

4. The correlation between variables. This is a more analytically advanced model to evaluate the impact of the different variables on the final ranking. Whether using Excel analysis Toolpack or SPSS, the user can interpret correlations between the goal rate from certain areas and the ranking. A demonstration is provided in the voice-over PowerPoint presentation.

​

II/ FURTHER ANALYSIS

The analysis can be extended, and it can be more generally improved through a few methods:

• Data cleansing ensures a better quality of data. Checks conducted on the available data have raised incoherencies (e.g.: more goals than shots). It is useful to build a scatter plot to spot outliers and to keep them out of the sample to avoid analysis bias.

• Data collection over a broader time period would enable trend analysis over time. A figure makes more sense when it is compared to other periods. This would enable to avoid one-off value due to a bad season, for example.

• Other aspects of netball need exploring, in order to get a wider range of information.

• Individual performances should be measured. Cognitive and physical data have to be combined to obtain the “big picture” of each of the players. Injuries and team composition over the season should also be taken into consideration.

• Team performances about court coverage turnovers or possession rate would also be useful to refine the strategy. This data can be modelled with heat maps for instance, which are very useful for location visualisation.