One technique to analyse game telemetry data is data clustering,which is the process of grouping data into small clusters. Eachcluster groups similar data which are distinct from one other. [10],[11].Clustering is a method that uses unsupervised neural networks,networks that have no need to know the desired output, learningwith input data only [12],[13].For this experiment, k-means clustering method was usedbecause there are other studies using this method to classify andgroup types of players [7], and even outside of gaming literaturethis method has been used to cluster data [17]. K-means is analgorithm where the number of clusters (k) are chosen and eachcluster centroid is initialized in a distinct place of the dataset. Afterthe initialization, centroids are iterated and, based on the Euclidiandistance between data and the cluster’s mean, centroids move andstart grouping the data into clusters until there are no movementsneeded and the clusters are set [10],[11]. This can be seen in Figure