The purpose of this thesis is to create 3D visualization methods for portraying the chemical cartography of the Milky Way using Gaia and GALAH data within the open-source software OpenSpace. While chemical cartography is astrophysics provides valuable insights into star formation and their kinematic evolution, existing maps predominantly remain static and two-dimensional, often lacking interactive analysis tools. This work enhances the existing point rendering pipeline, thereby enabling interactive analyses catering to both scientific research and public outreach.

To realize this vision, a more generalized approach to data ingestion was adopted. This involved extending the file reading capabilities to include CSV format, effectively eliminating the constraints imposed by the previous riding data formats. These changes not only enable the software to accommodate a broader range of data parameters but also introduce a user-friendly interface for seamless data selection and real-time range specification during the visualization process. As a result, adjustments to the GPU data layout were made to accommodate the flexible rendering technique. Moreover, enhancements in visual quality were achieved by removing invalid data values in the datasets, as well as updating the distance-based luminosity scaling of starts. In addition, comprehensive performance testing was conducted to evaluate the modification to the pipeline, the testing involved various aspects, including read/write speeds, load times, frame rates, and constructing internal data structures. Despite the challenges posed by larger datasets and an increased parameter count, the impact on runtime performance remained largely unaffected. However, initial load times were primarily negatively affected by larger datasets but can be mitigated by pre-generating the data structures. Furthermore, three-dimensional data representation techniques were explored in an attempt to enhance the visualization of star clusters and structures within the Milky Way datasets. This included a ray-casting volumetric visualization approach using two-dimensional transfer functions, leading to the development of an innovative volume rendering pipeline. An automated process was implemented for generating data-driven 2D transfer functions based on the underlying 2D histogram of selected data axes to facilitate ease of use. Moreover, the volume is automatically generated, and the resulting volume can be dynamically adjusted, offering options to vary resolution and apply two different data filtering techniques to refine the visualization. However, volumetric rendering introduces challenges, particularly in constructing effective transfer functions that yield meaningful visualizations. The quality of visual output is influenced by factors such as resolution, filtering method, and data characteristics. The relationships between these parameters are intricate, often requiring a trial-and-error approach to attain optimal results. 

In conclusion, this project presents a substantial advancement over the previous rendering pipeline, offering improved data handling and visualization capabilities. The integration of dynamic rendering and volumetric representation enables seamless storytelling and chemistry visualization, enhancing the user experience. In addition, comparisons between point and volumetric rendering underline their complementary roles: point rendering excels with sparse datasets and single parameters, while volumetric rendering offers advantages in scenarios involving spatially dense structures and multi-dimensional data. The refined pipeline facilitates easy adaption for researchers to visualize various datasets and empowers them to explore complex multi-dimensional data. Overall, this project has laid a foundation for further advancements in chemical cartography visualization and data exploration within the OpenSpace platform.