Implementation of Digital Twins, particularly Data-Driven Digital Twins, is transforming the modern manufacturing landscape by facilitating enhanced simulation, optimization and predictive capabilities. This paper investigates the current literature related to Data-Driven Digital Twins in Manufacturing context, tracing its foundations, technologies and applications. A critical analysis of the operational benefits of Data-Driven Digital Twins is conducted through a systematic review of existing publications. Various gaps identified, including the lack of explainability and effective integration of human expertise within Data-Driven Digital Twins, where current models often produce outputs that are difficult for stakeholders to understand, thereby hindering their widespread usability. To address this, this paper proposes a future research direction that explores the effectiveness of an interactive digital interface in improving the insights obtained from data-driven modelling methods embedded within Digital Twin systems.