Summary: Basics of Game Data Science

Takeaways and important terms

This chapter provided an overview of game data science and contextualized the field from a historical perspective to understand the evolution of methods and techniques presented in the course. Further, we examined how game data science generates new knowledge and insights following the scientific method. Finally, we presented a glossary to disambiguate nebulous terminology and foreshadow key concepts that will be recurring throughout the course. Key terms are repeated here.

Game data science

Game data science is the process of applying methods, practices, and techniques from existing relevant fields such as data science, business intelligence, geographic information systems, or social network analysis to the specific context of games for the purpose of informing and supporting decision-making.

Game analytics

Game analytics is historically related to business intelligence, and hence, it represents a subset of the methods and techniques that collectively form game data science.

Game telemetry

Game telemetry is data logged from game clients and/or servers.

Behavioral telemetry

Behavioral telemetry is the specific type of game telemetry that deals with the second-by-second behavior of players and the interaction between players and the game.

Knowledge discovery process

The knowledge discovery process is the process of finding actionable knowledge from data. It’s achieved using data mining methods (algorithms and visualization tools) to extract nontrivial knowledge from a large amount of data.

Inductive reasoning

Inductive reasoning is the method of deriving general principles from specific observations, often called bottom-up logic.

Deductive reasoning

Deductive reasoning is the process of reaching a certain conclusion by analyzing evidence refuting or solidifying a hypothesis, often called top-down logic. Game metrics are interpretable quantitative measures of in-game attributes; a metric always implies some level of abstraction from raw data.

Key performance indicators (KPIs)

KPIs are special game metrics that have been strategically selected to demonstrate how effectively a player can achieve an objective.

Community metrics

Community metrics are game metrics that deal with the social dimension of players at all resolutions, from forum activity to a number of friends or number of groups created or joined.

Customer metrics

Customer metrics are game metrics that deal with the user as a customer, taking into consideration all forms of financial transactions.

Gameplay metrics

Gameplay metrics are metrics that deal with the actual behavior of a player within a game.

Acquisition metrics

Acquisition metrics are a set of KPIs focused on new players and how they become engaged with a game. Retention metrics are a set of KPIs focused on maintaining active players by assessing how many players remain active within a game for a certain period. Progression metrics are a set of KPIs focused on the progress of players through a game.

Player profiling

Player profiling is the practice of dividing and grouping players into sets that are similar in specific ways. Segmenting is the process of assigning players to preexisting groups that were derived from both qualitative and quantitative research.

Clustering

Clustering is the process of finding similarities and differences among players based on objective, quantifiable measures through statistical methods.

Player modeling

Player modeling is the process of creating computational descriptors that can account for differences and similarities in players and predict the value of unseen data based on currently available data.