Assessing and predicting the complex concept of software quality is still challenging in practice as well as research. Activity-based quality models break down this complex con- cept into more concrete definitions, more precisely facts about the system, process and environment and their impact on ac- tivities performed on and with the system. However, these models lack an operationalisation that allows to use them in assessment and prediction of quality. Bayesian Networks (BN) have been shown to be a viable means for assessment and prediction incorporating variables with uncertainty. This paper describes how activity-based quality models can be used to derive BN models for quality assessment and pre- diction. The proposed approach is demonstrated in a proof of concept using publicly available data.