Marcelo Spiezzi Raimbault
Production
«
Behavioral Clustering in Open World Sandbox Games
Data analysis is one of the areas I am
techniques such as k-means and
the analysis of player data can highlight
data. I identified seven unique behaviors
passionate about. It is inspiring how
possible problems and provide a better
understanding of their behaviors. This
project was made in collaboration with
Nerd Kingdom and their game TUG over
an eight-month period, during which we
collected more than 400 distinct sessions
to study.
that showed possible gameplay styles
in an open world sandbox game, and
highlighted potential design issues with
TUG. The project results presented how
the understanding of behaviors can lead
to better decisions for the design and
production disciplines.
This thesis presented various challenges
This work was supported in part by
a custom mod to collect data in Lua,
15149094 as a part of the Brazil’s Science
throughout its execution, such as writing
managing and engaging the community,
building a data analyzer in C# to study
the data, and applying different clustering
46
Simplex Volume Maximization to the
PRODUCTION
the CAPES Foundation under process
without Borders Scholarship Program.