Project: Analysis of outWord Player Data
In order to decide what changes to make to outWord, my company’s first game, I undertook an in-depth study of the player data set. This required devising formal questions answerable with data we collected from player actions, investigating months of database snapshots of this data, crafting queries that extracted information relevant to these questions, and using statistical software to analyze and visualize the data to discover patterns in player behavior. These explorations revealed several important findings and validated some theories I had about the game, including pointing to potential usability issues and problems with game mechanics and suggested future improvements to make the game more engaging. Additionally, this exploration gave me a great excuse to learn how to use a sophisticated data graphics toolset in statistical software.
One interesting finding from the data set revealed that players differ in their walking behavior. I discovered that the vast majority of words were made covering greater distances at too great a speed for walking. This validated what we had heard in user-interviews: players often rode bicycles, took public transit or traveled in cars while playing. Check out some highlights from the interrogation in this slideshow, or check the entirety of the analysis on my project blog.
Skills: Information Visualization, Data Munging, Database Analysis, Information Analysis with R