Importance of Metrics to Game Data Science

Applications of metrics to game data science

As we can see, with such an extensive list of metrics, we can do many things, from modeling player behaviors, profiling players, predicting win/loss, or predicting churn, to mention a few examples. Let’s take player profiling as an example. Given the metrics in the preceding text, it’s not a surprise that different groups of players react differently to the same stimulus and behave very differently in the same game. It’s important to understand how players differ and in what way they’re similar because a one-size-fits-all approach to development is not often effective.

Player profiling is the practice of dividing players into similar groups in specific ways relevant to different stakeholders, such as brands, live operations, or designers. The similarities can be based on the following attributes:

• Socio-demographics: This includes age, gender, education, occupation, income, marital status, ethnicity, language, and religion.
• Geography: This includes country and urban/rural areas.
• Psychographics: This includes motivations, personality, needs, values, attitudes, and interests.
• Behavior: This includes spending habits, frequency, consistency, and duration of playing behavior.