Visual monitoring tool of association patterns in binarydata flows

Data streams are one of the most relevant new data sources, they refer to flows of data that come at a very high rate. The distinguishing feature in data streams analysis is that the focus is on transient relations. The present paper proposes a quantitative transformation of the binary attributes exploiting Multidimensional Correspondence Analysis (MCA) to describe the evolving association structures among attributes over different time-frames. The quantitative coding of the original attributes, in addition to synthesize information, make possible visualizations for different purposes, such as factorial maps, parallel co-ordinates and dendrograms.