Game analytics using natural language processing
A processor executes program code that represents a portion of a video game and adds a sequence of text strings that represent game events to a text log during execution of the program code. The processor (or another processor that has access to the text log) performs a natural language processing (NLP) analysis of the text log to determine one or more characteristics of the portion of the video game. In some cases, the NLP analysis includes a sentiment analysis that attempts to determine characteristics of a player's experience while playing the video game, summarization technology that creates a human-readable summary of an aspect of the game or a portion of the video game, a semantic NLP ML algorithm in the semantic similarity modality to answer questions regarding the player's experience during the video game, or grouping players in a multiplayer game based on in-game behavior.
1. A computer-implemented method for implementation by at least one processor, the method comprising:
executing program code that represents a portion of an application;
adding a sequence of text strings that represent events application to a text log during execution of the program code;
generating, based on a captioning model, a natural language description of at least one of a scene in the application and an interaction between entities in the scene;
adding the natural language description to the text log;
executing a natural language processing (NLP) analysis algorithm to perform an NLP analysis of the text log to determine at least one characteristic of the portion of the application, wherein performing the NLP analysis of the text log includes performing a sentiment analysis of the text log to determine characteristics of a user's experience while using the application; and
modifying program code that represents at least one of the NLP analysis algorithm or the portion of the application based on the analysis of the text log.
2. The computer-implemented method of claim 1 , further comprising:
using an application programming interface (API) to associate the events with text strings including at least one of numbers, letters, words, or phrases.
3. The computer-implemented method of claim 1 , wherein adding the sequence of text strings to the text log comprises adding the sequence of text strings to the text log in response to a director generating the events, wherein the director is implemented using a portion of the program code.
4. The computer-implemented method of claim 1 , wherein the application is a videogame and wherein executing the program code comprises executing a character behavior engine that uses a semantic NLP machine learning (ML) algorithm to generate a response to an input that represents a user action.
5. The computer-implemented method of claim 4 , wherein adding the sequence of text strings to the text log comprises adding a first text string associated with the input to the text log and adding a second text string associated with the response to the text log in response to the character behavior engine generating the response.
6. The computer-implemented method of claim 1 , further comprising:
generating, using a generative grammar, natural language statements based on information produced by a conventional analytics system; and
adding the natural language statements to the text log.
7. The computer-implemented method of claim 1 , wherein the application is a video game and wherein generating the natural language description comprises using a plurality of instances of the captioning model are used to generate the natural language description of the scene viewed from a plurality of character perspectives or arbitrary perspectives.
8. The computer-implemented method of claim 1 , wherein performing the NLP analysis to determine the at least one characteristic comprises generating a human-readable summary of an aspect of the application or a portion of the application based on the text log.
9. The computer-implemented method of claim 1 , wherein performing the NLP analysis to determine at least one characteristic comprises answering questions regarding the user's experience during the application by applying a semantic NLP ML algorithm in a semantic similarity modality based on the text log.
10. The computer-implemented method of claim 1 , wherein performing the NLP analysis to determine at least one characteristic comprises grouping players in a multiplayer video game based on in-game behavior of the players that is inferred from the text log.
11. The computer-implemented method of claim 1 , wherein performing the NLP analysis comprises constructing an experience curve based on the text log.
12. The computer-implemented method of claim 1 , further comprising:
comparing the at least one characteristic of the application to results of beta testing or focus groups for the application; and
defining at least one rule to modify the NLP analysis algorithm based on the comparing.
13. A non-transitory computer readable medium embodying a set of executable instructions, the set of executable instructions to manipulate at least one processor to perform the method of claim 1 .
14. An apparatus comprising:
a memory configured to store a text log comprising a sequence of text strings that represent events that occurred during execution of program code that represents a portion of an application; and
at least one processor configured to:
generate, based on a captioning model, a natural language description of at least one of a scene in the application and an interaction between entities in the scene;
add the natural language description to the text log;
perform a natural language processing (NLP) analysis of the text log to determine at least one characteristic of the portion of the application by:
executing an NLP analysis algorithm; and
performing a sentiment analysis of the text log to determine characteristics of a user's experience while using the application; and
modify program code that represents at least one of the NLP analysis algorithm or the portion of the application based on the analysis of the text log.
15. The apparatus of claim 14 , wherein the processor is configured to support an application programming interface (API) for associating the events with text strings including at least one of numbers, letters, words, or phrases.
16. The apparatus of claim 14 , wherein the text strings are added to the log in response to a director generating the events, wherein the director is implemented using a portion of the program code.
17. The apparatus of claim 14 , wherein the application is a video game and where processor is configured to execute the program code, and wherein the program code represents a character behavior engine that uses a semantic NLP machine learning (ML) algorithm to generate a response to an input that represents a user action.
18. The apparatus of claim 17 , wherein the processor is configured to add a first text string associated with the input to the text log, and wherein the processor is configured to add a second text string associated with the response to the text log in response to the character behavior engine generating the response.
19. The apparatus of claim 14 , wherein the processor is configured to:
generate, using a generative grammar, natural language statements based on information produced by a conventional analytics system; and
add the natural language statements to the text log.
20. The apparatus of claim 14 , wherein the application is a video game and wherein processor is configured to execute a plurality of instances of the captioning model are used to generate the natural language description of the scene viewed from a plurality of character perspectives or arbitrary perspectives.
21. The apparatus of claim 14 , wherein the processor is configured to generate a human-readable summary of an aspect of the application or a portion of the application based on the text log.
22. The apparatus of claim 14 , wherein the processor is configured to track meaningful concepts regarding the user's experience during the application by applying a semantic NLP ML algorithm in a semantic similarity modality based on the text log.
23. The apparatus of claim 14 , wherein the processor is configured to group players in a multiplayer video game based on in-game behavior of the players that is inferred from the text log.
24. The apparatus of claim 14 , wherein the processor is configured to construct an experience curve based on the text log.
25. The apparatus of claim 14 , wherein the processor is configured to:
compare the at least one characteristic of the application to results of beta testing or focus groups for the application; and
define at least one rule to modify the NLP analysis algorithm based on the comparing.