As Above, So Below

Exploring the known universe with data visualization.

Image by MidJourney (v4).

For centuries, humans have observed the skies and endeavored to understand the universe. The early two-dimensional charts and drawings of the observed universe were an attempt to retain that knowledge. They also helped early astronomers to execute the intricate calculations that led us to findings such as the trajectories of planets.

Copernicus’s model of the solar system was published in 1543. Image by Gordon Johnson from Pixabay.

From early charts that used visualization to chart our solar system to modern visualizations that use the abundance of data now available to render realistic 2D and 3D representations of galaxies and black holes, astronomy and data visualization have always worked hand-in-hand to enrich our knowledge of the universe.

As technology has advanced, the fields of astronomy and visualization have both made strides. Many advanced tools and techniques in data visualization space are currently available, but they are not always suitable for the purposes of astronomy researchers. The challenge for researchers in the field of astronomy is not just the volume of data, but the diversity of data that they deal with.

Astronomers and astrophysicists use various methods to collect their data, like observatories, space probes, and simulations. Depending on their area of interest, they could be collecting data for extragalactic, galactic, planetary, or solar system research.

To understand the broad and complex nature of astronomy research, we can think of the tasks the researchers perform in terms of these categories:

• Data wrangling: Transforming the gathered data into formats that can be used with the available visualization tools.
• Data exploration: Exploring the gathered data to form hypotheses, or just look for patterns that could be analyzed and will potentially lead to interesting findings.

Density map generated using the tool TOPCAT which lets researchers explore data in a visual form.

• Feature identification: Assessment of visualized data to identify features, which could be anything like clusters of galaxies or dark matter.
• Object reconstruction: Using images and observed data to create 2D or 3D representations of astronomical objects.

3D model of a nebula constructed using a 2D image as reference.

• Education and Outreach: Presenting astronomy research in a way that is accessible to a wider audience through spaces like museums and science centers.

Presentation on the surface of Mars at the Brilliant Minds conference.

To perform all these tasks, researchers use various techniques:

1. 2D/3D plots like pie charts, bar and line plots, histograms, scatterplots, etc.
2. Image processing to generate 2D images of data.
3. 3D rendering to generate representations of data of interest.
4. Interactive visualizations that allow filtering and the ability to zoom into areas of interest.
5. Reducing high-dimensional data into low-dimensional data, which essentially means narrowing down the types of data collected from sources so that researchers can focus on some specific areas.
6. Reducing the noise or uncertainty present in data which can occur due to the large-scale collection of data which is often fraught with randomness.
7. Making astronomy data ready for use on newer display platforms like those using virtual reality.

Considering all the different kinds of research that astronomers or astrophysicists perform on their data, and the techniques generally utilized to achieve that, it is not difficult to imagine the variety of sophisticated tools they would need. Tools and platforms such as Blender, yt, OpenSpace, Cosmicflows, Gaia Sky, Glue, among many others specialize in some or the other areas or techniques and are often used together to conduct end-to-end research.

While many tools have been developed to support intricate data visualization, researchers encounter challenges while using them for astronomy research. One such challenge is that different tools require data in different formats, so they cannot be used seamlessly without additional efforts from the researchers to ensure that the data from one tool can fit into another. Having this problem on a cosmological scale is definitely a monumental challenge.

Other opportunities for improvements in this area of visualization include an improved ability to query data, better support for the discovery of interesting astronomical features, enhanced interactions for visualizations, and many more.

References

• Lan, F., Young, M., Anderson, L., Ynnerman, A., Bock, A., Borkin, M.A., Forbes, A.G., Kollmeier, J.A. and Wang, B. (2021), Visualization in Astrophysics: Developing New Methods, Discovering Our Universe, and Educating the Earth. Computer Graphics Forum, 40: 635–663. https://doi.org/10.1111/cgf.14332