The main purpose of this STSM is to develop statistical approaches to predict cryptocurrency prices using historical time series data as well as information collected via social media. In particular, we are going to use white-box methods, rather than black-box ML techniques, which are currently highly popular but are of difficult interpretability. Our approach will be based on graphical models and vine models, which are dependence models where associations between variables can be represented graphically. The visual representation of both models aids the interpretation and communication of results to a wide range of audiences, making the outcomes accessible to and applicable by a range of stakeholders.