6. Appendix

6.1. Appendix Overview

This appendix includes a guide for contributing to OpenGameData, and a general glossary of terms.

6.2. Contributing to OpenGameData

We are always looking for people to contribute their work back to the main OpenGameData project. If you have added new features or fixed issues in any of the OpenGameData tools, whether in a fork or local Git clone, we invite you to submit a pull request through the corresponding GitHub repository.

6.3. Glossary

This page contains some important terminology for OpenGameData. Refer here if you are uncertain what a particular term means.

Note, the list below uses “user” terminology to decribe the terms - in general, the terminology below applies equally to “sessions” and “populations” as well.

  • event: A record of some game event, stored as a single row in the database.

  • Event: A record of some game event, whose structure conforms to the OpenGameData Event standard.

  • Detector: A module of code that “detects” certain user behaviors, based on the user’s event stream, and inserts new “generated” events into the stream. In OpenGameData, we write implementations of an abstract Detector class, where each implementation generates new events given event logs as input.

  • feature: Some metric or other piece of data, useful for analysis, that we can observe from gameplay event logs. For example, a feature could be a simple boolean representing whether “the player started level 1.” On the other hand, a feature could be complex, such as a string encoding the entire sequence of a gameplay session.

  • Feature: A record of some feature, whose structure conforms to the OpenGameData Feature standard.

  • Extractor: A code module that “extracts” Feature values from the user’s event stream, and returns that value to be associated with the user. In OpenGameData, we write implementations of an abstract Extractor class, where each implementation calculates a feature given event logs as input.

Features may be calculated at varying levels of analysis, described in the following definitions:

  • Aggregate Feature: A Feature that is calculated and yields one value across an entire session, user, or population event stream.

  • Unitized Feature: A Feature that is defined for some unit of gameplay within a game, such as levels, survey prompts, in-game quizzes, etc. Then there is one instance of the Feature for each instance of the unit. For example: A feature may be “clicks in the level”, with one instance for each of 20 in-game levels.

  • Per-Level Feature: An Unitized Feature that specifically has one instance for each level in a game. OpenGameData makes a special base module available to allow easier Feature engineering for games that include a notion of “level,” but its use is optional.

  • Population Feature: A Feature (whether aggregate or unitized) that is extracted from data across many sessions of gameplay. By default, all Features can be extracted at the session, player, and population levels; however, features that perform special accounting for multiple players (e.g. a feature that yields a distribution of values) can be marked as population-only.