The integration of Augmented Reality in various industries has seen a rise in the past couple of years. Today AR is used on a large scale for data visualization. It offers several benefits such as immersion in the data and improved presentation over the traditional methods such as using a desktop computer. In this master thesis, we focus on the impact that AR data visualizations have on the decision-making abilities of the user. We explore how the user’s understanding of the data is affected when it is presented in AR. Further, we also study the idea of having Data on the Go where the data follows the user at all times. This has the advantage of giving live updates to the user when they are on the move. We evaluate the performance of the users in taking a decision when they are observing the data in AR and at the same time completing a secondary task of walking. We adopt the use case of medical data for our visualizations. To prepare for this we conduct several interviews with medical staff to understand what kind of data is important for analysing the health of a patient. We also conduct an exploratory user study where the participants play the role of a doctor in a hospital. We present the participants with four unique use cases where they have to take important decisions about a patient’s health within a time limit. The data is presented to the participants in 2 formats: Immersive Condition (on the HoloLens) and Non-Immersive Condition (on the Computer). We compare these two conditions in terms of the accuracy of the decisions, the time needed and preferences of the different participants. Based on our user study we believe that the decision-making capabilities of a user remain unaffected irrespective of which condition they are in. For time-critical situations the user can swiftly reach a decision in both conditions and for non-emergency situations the time-based performance of the Immersive Condition is better as compared to the Non-Immersive Condition. Based on the subjective feedback we received, we believe that the Data on the Go scenario has a lot of potential if the movement and orientation of the data are improved.