Data Visualization aids in understanding the
Stellar Evolution of Binary Stars.
On The Visualization of Binary Stellar Evolution for POSYDON
We present BinaryStory, an open-source visualization tool for stellar binary evolution data. BinaryStory directly integrates with data, a departure from the conventional post-data processing practices for creating illustration graphics. It generates van den Heuvel diagrams based on output data from Binary Population Synthesis Code, POSYDON. BinaryStory creates a visual profile of the stellar evolution chain, supporting visual analysis of binary stellar models. Leveraging a mixture of illustrative and computational methods, researchers can gain a deeper understanding of their data while enhancing scientific communication and outreach.
. . .
Dean Kousiounelos, Seth Gossage, Vicky Kalogera
Introduction.
Telling the life story of binary stars is difficult and computationally expensive given the complexity of their evolution and the many permutations of possible stellar outcomes. In order to gain deeper insight into pathways of stellar evolution, we use the Binary Population Synthesis (BPS) code, POSYDON, to accurately simulate large populations of binary interactions and yield statistical properties without such a computationally expensive budget.
The inner structure of this code is an intermingled network of simulations such as the Modules for Experiments in Stellar Astrophysics (MESA) and algorithmic analysis of stellar evolution data (Fragos et al. 2023). This data can help many scientists understand how binary stars evolve. Furthermore, we begin to develop a story that describes how these binaries become compact objects (COs) and other extreme stellar objects; the progenitors of gravitational wave sources.
Despite having such a tool, fully grasping the binary stellar evolution story relies on interpreting data presentation. Hence, we employ visualization and illustrative methods to enhance data comprehension and convey the stellar narrative immediately upon data output. To aid in the analysis of data and relay the evolutionary narrative to the public, a visualization component must be developed in the POSYDON code.
Methods.
The development of our visualization tool involved a comprehensive approach, integrating both physical and computational materials. An essential goal was to create visualizations that not only facilitate scientific research for experts but also effectively communicate scientific concepts to the broader public. To achieve both of these objectives, the methods employed can be categorized into two main approaches: illustrative and computational. By combining the strengths of both approaches, we successfully established an intuitive and coherent visual framework that effectively merges the power of illustrative representations with the precision of computational visualization.
Illustrative Methods
Prior to computational work, initial illustrations were pivotal in creating the stellar evolution visualization tool. They depicted evolution stages and events while offering a flexible foundation for updates. Crafted with varying pencil hardness, archival ink pens, and color pencils in a muted palette, each detail was meticulously attended to. Spheres used compasses, and curves employed a flexible curve ruler. High-quality scans used 250 GSM white Stonehenge paper.
Early Illustrations of the Zero Age Main Sequence (ZAMS) and the Roche Lobe OverFlow (RLOF) stage. These illustrations (left) were completed with Derwent color pencils on 250 GSM Stonehenge paper and re-created as vector-like illustrations using Adobe Illustrator and Photoshop (right). Credit: Northwestern. Visualization by Dean Kousiounelos .
Adobe Illustrator converted scans to vector images, ensuring resolution quality and customization. Standardized dimensions guaranteed smooth computation for the final tool. Incorporating the essence of traditional visualizations like Van Den Heuvel diagrams, we aimed to develop a more comprehensive and detailed visual tool that is fit for POSYDON data. The layout facilitates a portrayal of binary star evolution, ensuring that researchers can easily grasp the relationships and transitions between different evolutionary stages in the evolutionary narrative. Stars are color-coded by temperature in the binary system, and their sizes are relative. We designed a minimalist and representative motif for accreting matter, accommodating the in-spiral of such matter for stages such as Roche lobe overflow (RLOF). A 3-dimensional perspective was applied to the Roche lobe, aiding in its dimensional understanding, and enabling future visualization coding. However, the spheres are kept nearly 2 dimensional to maintain the balance of informative but minimalist motifs present throughout the visualization.
Simplified potential diagrams (Tauris & van den Heuvel 2023), inspired by those in The Physics of Binary Star Evolution, were chosen for clarity, as the equipotential surfaces present in each event stage acted as a visual link between the RLOF event with its physical effect in the gravitational field of each star. The layout of the visualization adds visual dimensionality to the algorithmic diagram, Fig. 1 (Fragos et al. 2023). The Illustration of this tool, however, just represents a singular end result. BinaryStory must be able to be auto-generated with respect to any population run of the POSYDON code. Therefore, it adopts a computational contribution that allows the outcome of the visualization to be based on the POSYDON output data.
Computational Methods
After creating the visualizations, the next step was developing an interactive version using the POSYDON code. Leveraging Python packages such as Matplotlib, PIL, Cv2, we created an integrated Van Den Heuvel diagram gen- erator. First, a BPS model run, with specific keyword arguments, and a large population yielded a large Pandas DataFrame (DF). The function getstar returns one star from the population and displays its important information, such as state and event names, step names, mass, state, separation, etc., for the following visualization functions. Having established a DF, MiniVizView, with relevant stellar evolution information, we assign a function named getimage which receives the DF and extracts pertinent information for constructing a Van Den Heuvel diagram. Each chartobject acts as an image in the visualization and requires an axis and an image as input.
A new Van Den Heuvel diagram, with a timescale, that is generated directly from the population synthesis data from POSYDON. It depicts a zero-age main sequence binary (ZAMS) evolution to a merger, including the onset of Roche lobe overflow (RLOF), the common envelope stage (oCE1), as well as in-spiral and winds down steps before the merged stage. Credit: Northwestern. Visualization by Dean Kousiounelos.
The function vdh is the integral function that created the Van Den Heuvel Diagrams. It iterates over columns, searches for corresponding visualizations in a database based on a specific file path that is stitched together by the output of the getimage function, and places connected chart objects in chronological order. During this process, it uses PIL to turn the image into a numpy array of pixels with RGB values.
(Top)The RGB, HS, and Hex codes for the visualizations. These codes allow us to change the color of each star in the Van Den Heuvel diagram with an accurate color-to-effective temperature image. (bottom) Star1 is a 3000 K star and a specific RGB value is assigned to it. The hue is moved to be the same saturation as the star on the right so that the color scheme is balanced. Credit: Northwestern. Visualization by Dean Kousiounelos.
A list of effective temperatures of the stars at each stage is passed to a function that uses a database of corresponding Hexadecimal (hex) and RGB codes based on spectral type and temperature (Harre & Heller 2021). This is then used to color each star in the diagram. The visualizations, thus, hold the temperature and color information of the stars in the binary system. The function utilizes connectionpatches as the connecting pathways between each stage/event. The function accounts for null results and provides researchers with appropriate messages to notify if a visualization doesn’t exist or isn’t found. At each time step, points are plotted on a vertical line adjacent to the chart, introducing a temporal dimension to the visualization. The vdh function facilitates the creation of a well-organized timeline, documenting the stellar evolution of each binary system based solely on the output data from POSYDON.
Scientific Communication
BinaryStory plays a pivotal role in advancing scientific communication regarding the complex processes of stellar evolution. Through its visual representations, it effectively conveys information as an informative graphic, distinct from editorial illustration (Christiansen 2021). Striking a balance between raw data display and illustrative elements, BinaryStory deliberately adopts an illustrative approach to engage viewers with the data content. The visual appeal of BinaryStory lies in its simplicity, adhering to Tufte's design principles, such as data-to-ink ratio (Tufte 2016). Thoughtfully curated color aesthetics avoid high-contrast hues and discordant palettes, ensuring a coherent and visually captivating presentation. BinaryStory directly integrates with data, departing from conventional post-processing methods for graphics. It swiftly saves visuals in formats like png, jpg, or pdf, perfect for posters, articles, and presentations, enhancing the binary stellar evolution narrative. The resolution quality of the visualization enables easy import to graphics editors like Adobe Illustrator, allowing customization for scientific communication. As a valuable tool, BinaryStory facilitates clear and captivating depiction of binary star evolution across diverse formats, providing astronomers with the means of illustrating the compelling story of binary star systems.
Conclusion and Further Work
BinaryStory is a user-friendly, data integrated, tool that enhances comprehension of binary star systems by bridging complex data and illustrative elements for effective scientific communication. Its versatility and accessibility are valuable for researchers conveying the narrative of binary stellar evolution. Future developments could include interactive 3D visuals to further enhance user experience. Our study focused on one indexed star's evolution, but specialized 2-D Grid plots could elucidate channel contributions and simulation accuracy for the entire population. Beyond visualization, BinaryStory transforms communication by merging data analysis with storytelling to deepen understanding of binary stellar evolution.
References
Christiansen, J. 2021, Visualizing science: Illustration and beyond, , , accessed: 2023-8-9
Fragos, T., Andrews, J. J., Bavera, S. S., et al. 2023, ApJS, 264, 45
Harre, J.-V., & Heller, R. 2021, Astron. Nachr.
Tauris, T. M., & van den Heuvel, E. P. J. 2023, Physics of Binary Star Evolution. From Stars to X-ray Binaries and Gravitational Wave Sources, doi:10.48550/arXiv.2305.09388
Tufte, E. 2016, in Diagrammatik-Reader, ed. B. Schneider, C. Ernst, & J. W ̈opking (Berlin, Boston: De Gruyter), 219–230
Presented at the 243rd American Astronomical Society Convention, New Orleans
About Me
IMy name is Dean kousiounelos and I'm a rising junior at Lake Forest College double majoring in Physics and Art. This summer I had the great opportunity to work with Dr. Seth Gossage and Dr. Vicky Kalogera on a visualization code for binary population synthesis data from POSYDON.
I am working in the nexus of art and science, looking to bridge a gap that has long been defined by both artists and scientists alike through data visualization and graphic illustration. By creating visualizations that are integrated with the POSYDON data, a stellar story emerges, bridging both the quantitative and the qualitative aspects of the evolutionary narrative
Acknowledgements
I want to thank Dr. Seth Gossage and Dr. Vicky Kalogera for their mentorship; the REU leaders, Dr. Aaron Geller and Dr. Emily Leiner, and my cohort for their valuable support. This material is based upon work supported by the National Science Foundation under grant No.AST 2149425, a Research Experience s for Undergraduates (REU) grant awarded to CIERA at Northwestern University. Any opinions, findings, conclusions, or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. This research was supported in part through the computational resources and staff contributions provided for the Quest high-performance computing facility at Northwestern University which is jointly supported by the Office of the Provost, the Office for Research, and Northwestern University Information Technology.
