ABSTRACT
This chapter presents the emerging field of esports analytics by outlining replicable methods for analyzing character roles and team compositions using a data-driven approach. Video game genres such as multiplayer online battle arena, first-person shooter, and real-time strategy have started an era of electronic sports (esports) that has gained further ground over the past decade. Business Insider reports that esports are expected to bring in revenues of 1.5 billion dollars by 2020. The chapter aims to secure data of interest from Heroes of the Storm (HotS) matches. HotsApi provides a set of queries to acquire metadata about the replay files it hosts. Cluster analysis is an unsupervised machine learning method commonly used in exploratory analyses to group “similar” items within a dataset based on certain numeric attributes. The HotsApi documentation lists many different ways to acquire information about the matches hosted on the site.
