We present a web-platform for building, customizing and exploring phylomemies, introduced by [1, 2], for the synthetic representation of the evolution of scientific topics appearing in large document corpora. Phylomemies are built using computationally expensive data-centric workflows composed of complex text and graph mining steps. For producing representative and high-quality phylomemies experts generally have to compute several instances using different parameters until they get an optimal result. Our prototype is focused on increasing interactivity and flexibility in the usage of these workflows by introducing parallelization and incremental computation. The implementation is based on a scalable parallel architecture and the demonstration will illustrate the interactive creation of phylomemies on a collection with over 20 million references to scientific publications.