Introduction course scout/analyst: Part VI — Data context

Data context. People who know me, know that I’m a huge data advocate and lover. Working with data, making data profiles and analyses — that’s what makes me truly happy. But little did prepare me for the sensitivities and difficulties when working with data.

That’s why I want to look a bit closer to using data within the context. You will find similarities in some areas with preparing data, but here we look at what the context can mean in player statistics.

Last week we looked into reports and video analysis: how to use your video analysis in reports and how they can show different things. You can read it here.

Data and context

Data analysis is the process of examining data using statistical, mathematical, and computational methods to extract insights and make informed decisions. It involves collecting, processing, and interpreting data to draw meaningful conclusions that can help inform business decisions, public policy, and scientific research, among other applications.

However, data analysis is not a straightforward process. Data can be complex, messy, and noisy, and there are often many ways to interpret it. To make sense of the data, it is necessary to consider the context in which it was collected and the purpose for which it is being analyzed.

Context is the set of circumstances or conditions in which something exists or occurs, and it plays a critical role in data analysis. In data analysis, context refers to the background information, assumptions, and expectations that shape the interpretation and analysis of data. Without context, data analysis can be misleading or incomplete, leading to incorrect conclusions and poor decision-making.

There are several reasons why context is important in data analysis:

Understanding the purpose: Context helps to identify the purpose and objectives of the analysis. Knowing what questions need to be answered or problems need to be solved can help guide the analysis and ensure that it is focused and relevant. For example, a company might want to analyse customer data to identify trends and preferences to inform marketing and product development strategies. The context of this analysis would include understanding the company’s goals and priorities, the competitive landscape, and the target market.

Identifying biases: Context helps to identify any biases or assumptions that may be present in the data or analysis. These biases can skew the results and lead to incorrect conclusions if not properly accounted for. For example, a study on the health effects of a particular drug might have been funded by the drug manufacturer, creating a conflict of interest. The context of this study would include understanding the funding source and any potential biases or conflicts of interest.

Interpreting results: Context provides a framework for interpreting the results of the analysis. For example, understanding the historical or cultural context of the data can help to explain certain trends or patterns that may not be immediately apparent. In addition, context can help to identify outliers or anomalies that may require further investigation. For example, if a company’s sales data shows a sudden spike in one region, the context of this data might include understanding recent marketing campaigns or local events that could have contributed to the increase in sales.

Communicating results: Context helps to make the analysis more accessible and understandable to others. By providing context, it is easier to explain the significance of the results and how they relate to the larger picture. For example, a scientist might need to explain the results of a research study to policymakers or the general public. The context of this communication would include understanding the audience’s knowledge and expectations, as well as any potential implications or limitations of the study.

In addition to these reasons, context is also important because it helps to ensure that data analysis is ethical and socially responsible. By considering the context of the data and analysis, it is possible to identify potential harms or unintended consequences and take steps to mitigate them. For example, a company might collect data on its employees’ productivity, but without proper context and safeguards, this data could be used to unfairly evaluate or discipline employees.

Relating to football

In football, data analysis is increasingly being used to gain insights into player performance, team tactics, and game outcomes.

For example, a football team might use data analysis to analyse their opponents’ playing style and identify potential weaknesses to exploit in upcoming matches. Context is crucial in this analysis because the team needs to understand the broader context in which their opponents are playing, such as their recent form, injuries, and tactical approach. Without this context, the analysis could be incomplete or inaccurate, leading to ineffective game plans.

Similarly, data analysis can be used to evaluate a football player’s performance, such as their passing accuracy or shooting efficiency. However, the context of the data is critical in understanding the player’s performance. For instance, a player might have a high number of successful passes, but this might not be significant if these passes are made in less important areas of the field or do not contribute to creating scoring opportunities. In contrast, a player who makes fewer passes but with greater impact in creating chances may be more valuable to the team.

Context is also essential in interpreting the results of football games. For example, a team might lose a match, but if the context of the game is understood, it might be possible to identify reasons for the loss that do not reflect on the team’s quality. For instance, a team might have faced a particularly challenging opponent, or several key players might have been injured or absent. Understanding the context of the game allows for a more accurate assessment of the team’s performance and the factors that contributed to the result.

Data context: Player

Often we get data from a player and without translating or evaluating it, it can be used for anything. Often it works well on social media to get some traction, but that doesn’t mean it will mean good analysis. A good example of this is percentage of accurate passes. It’s often said how good player x is in passing because of his pass completion, but what does it actually mean?

While pass completion rate is a useful statistic to analyse a player’s performance in footba, it may not always be the best statistic to judge a player on. Here are a few reasons why:

1. Role and position: A player’s role and position on the field can greatly affect their pass completion rate. For example, a central midfielder who is responsible for distributing the ball to teammates in advanced positions may have a higher pass completion rate than a forward who is expected to take more risks and attempt more difficult passes. Therefore, a high pass completion rate does not necessarily mean that a player is performing well overall.
2. Pass difficulty: Pass completion rate does not take into account the difficulty of the passes that a player attempts. A player who only attempts safe, short passes may have a high pass completion rate but may not be contributing much to the team’s attacking play. Conversely, a player who attempts more difficult, creative passes that may have a lower completion rate may be more valuable to the team’s attacking play.
3. Style of play: Some teams play a possession-based style of football where a high pass completion rate is prioritised, while others prioritise direct, attacking play with fewer passes. In these cases, a player’s pass completion rate may not accurately reflect their performance in the team’s style of play.
4. Other factors: There are other factors that can affect a player’s pass completion rate, such as the quality of their teammates, the opposition’s tactics, and the conditions of the pitch. Therefore, pass completion rate should be analysed in conjunction with other statistics and contextual information to get a more complete understanding of a player’s performance.

The same applies for when we for example look at goalkeeper statistics used. One favourite of everyone is the number of saves. While the number of saves is an important statistic for goalkeepers, it may not always be the best statistic to judge a keeper on. Here are a few reasons why:

1. Quality of saves: The number of saves a goalkeeper makes does not necessarily reflect the quality of those saves. A goalkeeper may make a large number of saves but many of them could be routine saves that don’t require much skill, while a few crucial saves that require exceptional skill could be more important for the team’s overall performance.
2. Defense strength: The number of saves a goalkeeper makes can also be influenced by the strength of their defense. If the defense is weak and allows many shots on goal, the goalkeeper may have to make more saves but this doesn’t necessarily reflect their ability or performance. Conversely, a goalkeeper playing behind a strong defense may make fewer saves, but still perform at a high level.
3. Playing style: A goalkeeper’s playing style can also impact the number of saves they make. Some goalkeepers may prefer to come off their line to claim crosses and clearances, while others may stay closer to their line and make more saves from shots on goal. This can result in differences in the number of saves made and should be taken into consideration when evaluating a goalkeeper’s performance.
4. Other factors: There are other factors that can affect a goalkeeper’s performance, such as their distribution, decision-making, and ability to organise the defense. These factors may not be reflected in the number of saves a goalkeeper makes and should be analysed in conjunction with other statistics and contextual information to get a more complete understanding of a goalkeeper’s performance.

Final thoughts

Data context is very, very important. Without giving meaning and context to your data, your data is basically useless to use in a data analysis. In that light it’s always important to give it the proper context or have someone within your organisation who can. If you do that, you can benefit from it.

I have given just a couple of examples, but it’s the critical thinking that’s of vital importance here.

Next week we will come to the seventh part of the tutorial We will have a look at data visualisation.

Originally published at https://zonalpressing.com on March 15, 2023.