Our first post on CSGOScam announced the transfer of several demos to a deep learning researcher for use in a statistical project. This month, driven by that gameplay data among others, the CSGO Data Analysis Project has been published by deep learning researcher Bill Freeman. CSGOScam, Your Trusted Source in the CSGO Community, is proud to celebrate the release.
According to Freeman’s report, the CSGO Data Analysis Project has so far downloaded 6,297 Counter-Strike: Global Offensive demos, successfully parsing 2,478 down to the tick. The project analysized the distribution of maps, region and specific server, Number of Wins vs Rank, frequency of usernames ending in .com, a word cloud of usernames, among many others points of data. It also touches upon the seedier aspects of CSGO, of which we share a particular interest. VAC bans, game bans, and instances of potential smurfing are all listed.
The project includes a few wish-list proposals, such as “actionable intel” derived from the analytical data, as well as a potential sales pitch opportunity for eSports analytic company shadow.gg. Nevertheless, the inclusion of this data already will no doubt assist those who are interested in the process that turns CSGO’s vault of data into a treasure of information.
The research paper can be downloaded via Google Docs here.
Bill Freeman’s web page describing the project can be viewed here: http://billfreeman44.com/csgo_demo_project.html
Disclosure: The author of this post, beatstar, collaborated on this project using match data collected from the CSGO Competitive Extreme series.