As product lifecycle management is starting to be a standard information system in most companies, a data manipulation issue starts to emerge. Indeed, the information contained in a system often needs to be accessible or even transferred to another one to exploit a new software's functionalities. With the Big Data era, new methods of analysis start to emerge, and the graph visualisation of data enables a better human understanding of the data itself. We propose to address the PLM migration issues with a new approach based on analysis of information system and data topology. The data is passed onto a graph and bundles are made thanks to clustering algorithms; this bundling enables a better understanding of the data and ease migration. However, PLM has links between links which complicate the transition to graphs. Finally, the described use-case proves that data bundling eases the data migration and prevents some usual pitfalls and delays.