Having looked at a number of approaches of clustering, we introduce a different setting for clustering: data are not available in a full batch but need to be processed incrementally. We start with a discussion on the requirements for online clustering algorithms, and move onto the online averaging mechanism, introducing thereof the competitive learning algorithm and its variations, including the self-organizing maps, neural gas, and the leader–follower method. This is followed by an introduction of a few interesting applications of online clustering in various learning paradigms.