From Data Point to Destination

Abstract

The two companion papers describe a single system: a catalogue interface that renders every system in the Open Exoplanet Catalogue as an explorable three-dimensional scene, and the classification-driven procedural pipeline that gives each planet a parameter-derived appearance. This paper steps back from the one case to ask what it is an instance of. Exoplanets are the most pictured objects in modern science that have never been seen: three decades of discoveries have reached the public almost entirely through imagery invented on their behalf, from press-release artist's impressions to planetarium software to procedural games. We survey that landscape, the institutional visualisers (NASA's Eyes on Exoplanets and the Exoplanet Travel Bureau; ESO and ESA press imagery), the open planetarium platforms (Celestia, Stellarium, OpenSpace, Gaia Sky), and the procedural universes (Space Engine, No Man's Sky), against what is actually measured, and against the research on how astronomical imagery is received.

Two findings organise the synthesis. First, the axis that matters is not realism but disclosure: the systems that serve understanding best are those in which a viewer can discover which features are measured and which are invented, and the genre's recurring failure is the unlabelled blend. Second, interactivity changes the epistemics: a static image asserts, while an explorable model can expose its own parameters, and this, more than fidelity, is the medium's genuine advantage. We close with what generalises: a speculation-labelling discipline for scientific visualisation, the case for deterministic procedural imagery as reproducible and therefore criticisable illustration, and the limits of any picture as a substitute for understanding what a light curve is.

1. Introduction

In February 2017 the discovery of seven temperate, Earth-sized planets around TRAPPIST-1 (Gillon et al. 2017) was announced with a press kit whose imagery travelled further than any measurement in the paper: surface vistas of reddish rock under a swollen salmon sun, and a poster advertising the best habitable-zone vacation within twelve parsecs of Earth. The data behind the announcement were transit light curves, periodic dips in one star's brightness, and masses inferred from the timing perturbations the planets exert on one another. No property of any TRAPPIST-1 planet's surface has ever been observed. The imagery was labelled as artist's concept in the caption, as the genre's conventions require; the label did not travel with the image.

This episode is not a scandal; it is the normal condition of exoplanet communication, and it has been since 51 Pegasi b arrived in 1995 with no possible photograph (Mayor & Queloz 1995). The field's public existence is almost entirely visual, and almost entirely invented. What varies, and what this paper surveys, is how the invention is done, how it is disclosed, and what different choices cost and buy.

The companion papers describe one set of choices: a whole-catalogue interface whose planet imagery is derived from recorded parameters through a documented classification (companion paper 1, the system; companion paper 2, the appearance pipeline and its speculation gradient). This paper situates those choices in the landscape and asks the general questions: what can visualisation actually do for the public understanding of objects that cannot be seen, and where are the limits that no rendering choice removes?

Contributions

This paper is a synthesis; it reports no new system and no new measurements.

1. A map of the exoplanet-visualisation landscape across five families, press imagery, institutional interactives, open planetarium platforms, procedural universes, and parameter-derived rendering, compared on data fidelity, disclosure, interactivity and reproducibility.
2. An argument that disclosure, not realism, is the load-bearing axis, assembled from the imagery-reception literature and the survey.
3. A statement of what interactivity changes: an explorable model can surface its own evidential status in a way a static image cannot.
4. A statement of the limits: what no visualisation, however honest, can do, with the companion project located inside those limits rather than above them.

2. A vocabulary

Three distinctions recur and are worth fixing.

Index, illustration, simulation

Borrowing the documentary vocabulary: an index is causally produced by its object (a transit light curve, a direct-imaging point source, a phase-curve temperature map); an illustration is made by an agent to depict the object (every exoplanet surface view ever published); a simulation is generated by a model whose inputs include measurements (a climate model of a locked world; a parameter-derived procedural render). The public reflexively reads astronomical images as indexical, as photographs, because for a century the genre's famous images, Hubble deep fields, Apollo earthrises, were, however processed, indexical (Kessler 2012; Rector et al. 2007). Exoplanet imagery inherits that reading without the underlying causation, which is the genre's structural hazard.

The disclosure problem

"Artist's impression" is a caption-level label on an image-level artefact. Captions detach; images circulate. Any labelling regime that lives outside the pixels loses to reposting, a dynamic documented across misinformation research generally and visible to anyone who has met TRAPPIST-1 surface art without its caveat. Disclosure that survives circulation must be structural: either in the image itself, or carried by an interactive context the viewer cannot help but traverse.

The speculation gradient

Companion paper 2 grades every feature of a rendered planet as measured, derived, theory-guided, or fictive. The grading generalises to the whole genre and is used as this paper's comparative instrument: for each system surveyed, the question is not how good it looks but which tiers it shows, and whether the viewer knows which is which.

3. Three decades of planets, as the public saw them

A compressed history, keyed to what existed as evidence against what was shown.

1992–1995: planets with no possible face

The first confirmed exoplanets orbited a millisecond pulsar (Wolszczan & Frail 1992), lethal, lightless worlds announced essentially without imagery. 51 Peg b (Mayor & Queloz 1995) established the pattern that held for a decade: a wobble in a spectrum, communicated as a painterly gas giant. The evidence was one number-pair (period, m sin i); everything visible in the art was tier 4.

2000–2008: shape and colour arrive at the margins

The first transits (Charbonneau et al. 2000) made planets geometrical: a disc's area, if never its face. Thermal phase mapping put one number-map on a hot Jupiter (Knutson et al. 2007); HR 8799 delivered the first multi-planet direct image, four moving points of light (Marois et al. 2008), and press imagery duly rendered four banded giants no instrument had resolved.

2009–2015: the statistical era

Kepler (Borucki et al. 2010) transformed the field from specimens to demographics, thousands of candidates, the radius distribution, occurrence rates, and visualisation bifurcated: scatterplots for the population, ever-richer artist's impressions for the individuals. The Fulton gap and the mass–radius relations on which companion paper 2 builds its classification are products of this era; so is the Exoplanet Travel Bureau (JPL's Visions of the Future poster series, 2015–), which abandoned realism for knowing retro-futurist tourism, arguably the genre's most honest artefact, because no viewer mistakes a WPA-style poster for a photograph.

2016–2017: the nearest and the seven

Proxima b (Anglada-Escudé et al. 2016) and TRAPPIST-1 (Gillon et al. 2017) put temperate rocky worlds around the nearest stars, and produced the imagery discussed in §1, including, in TRAPPIST-1's case, coordinated art direction across agencies that fixed the system's public look, the "1e as ocean world" iconography, from data that constrained bulk density at best. The companion project's featured TRAPPIST-1 e renders as a tidally locked ice-ocean eyeball for stated, literature-cited reasons; companion paper 2's point stands, that this too is a tier-3 reading, one possible world among the permitted.

2018–present: molecules, maps, and JWST

TESS (Ricker et al. 2015) industrialised discovery; JWST moved characterisation from detection to inventory, carbon dioxide at WASP-39 b (JWST ERS Team 2023), spectra that distinguish clear from cloudy atmospheres (after Sing et al. 2016; Kreidberg et al. 2014). Evidence about appearance remains at the tier of a colour (Evans et al. 2013) and coarse brightness maps (Demory et al. 2016). The gap between what is measured and what is shown has narrowed slightly, and it remains vast.

4. The landscape of systems

Press imagery (ESO, ESA, NASA centres)

The highest craft and the widest reach; scientifically advised, caption-labelled, structurally undisclosed (§2). Its reception is the best-studied: the aesthetics-and-astronomy line of research finds that imagery drives engagement, and that the public's comprehension improves with explanatory scaffolding, captions, scale cues, expert framing, rather than with image fidelity itself (Smith et al. 2011; Rector et al. 2007 for the production ethics of observatory imagery). The finding transfers: what the audience needs attached to an exoplanet image is not more realism but more status information.

Institutional interactives

NASA's Eyes on Exoplanets renders the confirmed population as a navigable 3D catalogue tied to the NASA Exoplanet Archive (Akeson et al. 2013): authoritative data, stock appearance classes, implicit disclosure (a viewer may infer the planets are generic; nothing says so). Its virtue is the tie between position in the interface and record in the archive, the interactive equivalent of a citation.

Open planetarium platforms

Celestia (2001–) and Stellarium established open, extensible sky rendering; OpenSpace (Bock et al. 2020) and Gaia Sky (Sagristà et al. 2019) brought research-grade catalogue navigation, positional data at Gaia scale with presentation-quality contextualisation. Their centre of gravity is where things are, which is also where their data are strongest; planetary surfaces are textures, out of scope and mostly treated as such.

Procedural universes

Space Engine renders a seamless universe, procedurally extending real catalogues into the unmeasured; No Man's Sky (2016) generates worlds without any observational referent. Both demonstrate the aesthetic power of the technique the companion project borrows, and both, by design, erase the measured–invented boundary entirely: in Space Engine a real catalogued star and a procedural filler star present identically. As science communication they are cautionary in precisely the dimension in which they are triumphant as experiences.

Parameter-derived rendering (the companion project)

Every planet's appearance a documented function of its record, with the speculation gradient articulated feature by feature. Its weaknesses are the mirror of its commitments: no artist rescues an ugly or misleading procedural outcome, the disclosure currently lives in papers rather than in the interface, and its catalogue (the OEC) is community-maintained and ageing (companion paper 1, §6). It is presented here not as the genre's solution but as one point in the design space that the rest of the landscape leaves unoccupied: auditable invention.

The comparison, compressed:

5. What visualisation can do

Read against the reception research and the landscape, four capabilities are real and distinguishable.

Orientation

Interactive 3D genuinely communicates configuration: that TRAPPIST-1's system would fit inside Mercury's orbit; that hot Jupiters skim their stars; what a resonant chain looks like in motion. These are tier 1–2 facts, measured periods and derived geometry, the medium's home ground, and no static image or table communicates them as directly. The companion viewer's engineering (Keplerian propagation, true relative rates) spends its effort exactly here, deliberately.

Instantiation

A rendered world makes a category graspable: "tidally locked ice-ocean eyeball" is a phrase until one orbits it. The risk, that instantiation is read as portrait, and the mitigation, graded disclosure, are the subject of companion paper 2; the capability itself is genuine, and it is why the genre exists.

Scaffolded engagement

The reception literature's consistent finding: imagery recruits attention and interest across expertise levels, and understanding follows when explanation is attached (Smith et al. 2011). An interface that makes the explanation load-bearing, click the planet, see the record, the picture as the door to the data, converts the genre's hazard into its pedagogy.

Honest fiction, reproducibly

A deterministic procedural pipeline makes illustration criticisable in a way one-off art is not: the same record renders the same world for every viewer, the mapping is versioned code, and a domain expert can audit the function rather than debate a painting. This is a small, real epistemic upgrade, peer-reviewable illustration, and it is the companion project's central offer to the genre.

6. What visualisation cannot do

It cannot close the appearance gap

Nothing renderable today is a prediction of any specific planet's face, and, outside a future of resolved imaging that no funded instrument promises, nothing will make it one. The ceiling of companion paper 2, plausibility rather than prediction, is the genre's ceiling, not one project's.

It cannot label itself into viewers' memories

Structural disclosure improves on captions, but the reception research gives no licence for the belief that any regime makes speculation-status stick once images leave the interface. Screenshots strip context as surely as reposts strip captions. The honest statement is harm reduction, not solution.

It cannot substitute for the evidence's own form

The deepest understanding of the field's epistemics is understanding what a light curve and a spectrum are: that we know these worlds as shadows and wobbles. A visualisation that only shows worlds, however scrupulously graded, teaches the conclusions while hiding the reasoning. The genre's unclaimed territory, and this survey's most concrete recommendation, is the interface that renders the evidence alongside the invention: the actual transit depth behind this planet's radius, the actual radial-velocity curve behind its mass, so that the plausible world and the real measurement are one click apart. The companion project does not yet do this either; it is the roadmap item this paper exists to motivate.

It cannot outrun its catalogue

Every honest visualisation inherits its database's coverage, currency and biases; the transit surveys' geometric bias toward close-in planets skews any rendered population toward the very worlds, hot, locked, extreme, that look least like home. Companion paper 1's limitations section is one instance of a general law.

7. What generalises

For builders of scientific visualisation beyond this field, the survey compresses to four transferable disciplines.

1. Grade the speculation (measured, derived, theory-guided, fictive) as a design artefact, not a caveat, and put the grading in the interface, not the paper.
2. Make the picture the door to the data. Every rendered object should resolve, in one interaction, to the record that licensed it.
3. Prefer deterministic, versioned invention where invention is necessary: reproducible fiction can be audited, diffed and corrected; inspired fiction can only be replaced.
4. Record the projection. Scale factors, clamps and exaggerations are part of any legible model of astronomical space; keeping them explicit, in code and in the interface, makes the model reproducible and the distortion teachable.

And one deflationary through-line, the analogue of every honest survey's conclusion: the decisive factor is not the rendering technology. It is whether the maker treats the viewer as someone entitled to know what kind of statement each pixel is.

8. Conclusion

Exoplanets will remain invisible for the foreseeable future, and they will remain among the most pictured objects in science. That combination is not a problem to be solved but a condition to be managed, and the survey's finding is that it is managed well or badly along one axis above all: disclosure. The companion project's wager, a whole public catalogue rendered as worlds, every world a documented function of its record, with an interface that must eventually carry its own epistemics, is one way to manage it. The genre's history suggests the images will travel either way. What a builder controls is whether, when an image is traced back to its source, the source can show its receipts. The two companion papers are this project's receipts.

Acknowledgements

This paper synthesises the work of the observers, archivists, artists and engineers it surveys, above all the maintainers of the open catalogues (Rein 2012; Akeson et al. 2013; Schneider et al. 2011) on which the companion project and much of the landscape stand.

References

Discoveries

• Wolszczan, A., & Frail, D. A. (1992). A planetary system around the millisecond pulsar PSR1257+12. Nature 355, 145–147.
• Mayor, M., & Queloz, D. (1995). A Jupiter-mass companion to a solar-type star. Nature 378, 355–359.
• Charbonneau, D., Brown, T. M., Latham, D. W., & Mayor, M. (2000). Detection of Planetary Transits Across a Sun-like Star. ApJ Letters 529, L45–L48.
• Marois, C., et al. (2008). Direct Imaging of Multiple Planets Orbiting the Star HR 8799. Science 322, 1348–1352.
• Borucki, W. J., et al. (2010). Kepler Planet-Detection Mission: Introduction and First Results. Science 327, 977–980.
• Ricker, G. R., et al. (2015). Transiting Exoplanet Survey Satellite (TESS). Journal of Astronomical Telescopes, Instruments, and Systems 1, 014003.
• Anglada-Escudé, G., et al. (2016). A terrestrial planet candidate in a temperate orbit around Proxima Centauri. Nature 536, 437–440.
• Gillon, M., et al. (2017). Seven temperate terrestrial planets around the nearby ultracool dwarf star TRAPPIST-1. Nature 542, 456–460.

Characterisation (appearance-relevant; fuller list in companion paper 2)

• Knutson, H. A., et al. (2007). A map of the day–night contrast of the extrasolar planet HD 189733b. Nature 447, 183–186.
• Evans, T. M., et al. (2013). The Deep Blue Color of HD 189733b. ApJ Letters 772, L16.
• Kreidberg, L., et al. (2014). Clouds in the atmosphere of the super-Earth exoplanet GJ 1214b. Nature 505, 69–72.
• Sing, D. K., et al. (2016). A continuum from clear to cloudy hot-Jupiter exoplanets. Nature 529, 59–62.
• Demory, B.-O., et al. (2016). A map of the large day–night temperature gradient of a super-Earth exoplanet. Nature 532, 207–209.
• JWST Transiting Exoplanet Community ERS Team (2023). Identification of carbon dioxide in an exoplanet atmosphere. Nature 614, 649–652.

Data infrastructure

• Rein, H. (2012). A proposal for community driven and decentralized astronomical databases and the Open Exoplanet Catalogue. arXiv:1211.7121.
• Akeson, R. L., et al. (2013). The NASA Exoplanet Archive. PASP 125, 989.
• Schneider, J., et al. (2011). Defining and cataloging exoplanets: the exoplanet.eu database. A&A 532, A79.

Visualisation systems and imagery studies

• Rector, T. A., Levay, Z. G., Frattare, L. M., English, J., & Pu'uohau-Pummill, K. (2007). Image-processing techniques for the creation of presentation-quality astronomical images. AJ 133, 598–611.
• Smith, L. F., Smith, J. K., Arcand, K. K., Smith, R. K., Bookbinder, J., & Keach, K. (2011). Aesthetics and Astronomy: Studying the public's perception and understanding of imagery from space. Science Communication 33(2), 201–238.
• Kessler, E. A. (2012). Picturing the Cosmos: Hubble Space Telescope Images and the Astronomical Sublime. University of Minnesota Press.
• Bock, A., et al. (2020). OpenSpace: A System for Astrographics. IEEE TVCG 26(1), 633–642.
• Sagristà, A., Jordan, S., Müller, T., & Sadlo, F. (2019). Gaia Sky: Navigating the Gaia Catalog. IEEE TVCG 25(1), 1070–1079.
• NASA/JPL-Caltech. Eyes on Exoplanets; Exoplanet Travel Bureau / Visions of the Future poster series (2015–). https://eyes.nasa.gov; https://exoplanets.nasa.gov/alien-worlds/exoplanet-travel-bureau/
• Space Engine. https://spaceengine.org (V. Romanyuk, 2010–).
• Hello Games (2016). No Man's Sky.
• Celestia (2001–), https://celestiaproject.space; Stellarium, https://stellarium.org.

Appendix: Sourcing notes

1. Reception-research claims are drawn from the aesthetics-and-astronomy line (Smith et al. 2011 and successor studies) and from the image-production literature (Rector et al. 2007; Kessler 2012); the application of their findings, developed on indexical observatory imagery, to non-indexical exoplanet art is this paper's extrapolation, flagged as such.
2. Descriptions of proprietary systems (Space Engine, No Man's Sky, Eyes on Exoplanets) are from public documentation and direct use, not source audit; their internal data handling may differ from the observed behaviour described.
3. The TRAPPIST-1 press-imagery account describes the public artefacts (the February 2017 releases and the Travel Bureau poster) and asserts nothing about internal editorial process at the agencies involved.