Bridging UX and UR
Highlights:
UX Design - Gameplay data analysis
In-game Surveys
quick Likert-scale inquiries with ux in mind
For Lirica, I designed and implemented in-game survey systems used to create knowledge and insight on a player’s experience - without the necessity of an in-person playtest. Surveys are quick, and minimally disruptive to player engagement, while gathering self-reported feedback not typically accessible through a post-test questionnaire.
This is done by the placement of survey “snaps” at transitions of low intensity, and through the use of simple measurements, such as the three-point Likert scale. Snap Surveys became central to our user research pipeline as in-person playtests became scarce due to growing health concerns.
Gameplay Analysis
Exploring data from remote playtests
As we moved into our primary testing sprints, I visualized data gathered by the Telemetry systems for user research reporting. Tableau made it simple for me to create engaging visual aids for the team. The data was then presented to level and quest designers and used to drive creative decisions. The automated nature of the systems became crucial due to the loss of our ability to conduct in-person playtests during the Covid-19 quarantine.
Insights:
Quest Completion Order
Individual Quest Rating
Player Pathing
New/Returning Player
Player Drop-off
All sessions player drop-off (session ends)
Single session pathing
Technical Design and Systems
Analytics Links using Azure Playfab
Playfab is a leading Games LiveOps service from Microsoft. I used the Unreal Engine Marketplace Plugin to export and store our gameplay data.
Blueprints
Thanks to the extensive PlayFab documentation, much of the blueprint logic was straight-forward. I simply needed to make a few adjustments to accommodate the design of the telemetry systems.
Databases were separated by release and debug modes, allowing for differentiation between official playtesting, and developer trials.
PlayFab JSON Objects made creating modular data Key and Value pairs simple, especially when formatting data for easy wrangling.
Data Wrangling
I love python
Snippet from data_aggregate.py
I created a simple Python script to aggregate and format all individual session data files taken from our PlayFab Database. The data was appended to an ‘all_session’ files, ready for visualization in Tableau.
Given position data in world-space coordinates, the floats needed to be normalized and zero-based before moving into tableau.
Session ID’s were generated using the current system-clock time as a randomizer.