Decreasing network downtime for wireless telecommunication systems is becoming more and more important. Therefore, degrading system performance needs to be identified long before key performance indicators pick up on it. As a lot of information about what is going on in the telecommunication network can be found in counter readings, exploratory data analysis, such as unsupervised clustering algorithms and statistical methods, can be used to assist the telecommunication operator by grouping scenarios that appear to be similar. In this paper, we apply a clustering method known from the area of topic modeling and usually used to find out what topics are included in a collection of text documents. In our approach, counters are treated as words and a group of counter readings at a point in time is treated as a document. The results show that the method is capable of identifying statistical relationships between counter readings and could be used by telecommunication operators to investigate the relationships between counters; and hence, increase the operators' understanding of ongoing network traffic situations.