Finding Earth 2.0 is a reproducible computational astrophysics project asking a deliberately difficult question: among the exoplanets we have actually measured, which worlds most closely satisfy physically motivated conditions associated with an Earth-like, potentially habitable planet — and how strong is the evidence behind each one?
The project does not begin with a hand-picked list of famous planets. It begins with public astronomical archives, preserves provenance, propagates published uncertainty, and keeps Earth similarity, habitable-zone position, rocky plausibility, atmospheric observability and evidence for biology as separate questions. That distinction is central to the analysis.
01 — We Have Thousands of Planets. Almost None Are Earth 2.0.
The age of exoplanet discovery has transformed astronomy. Thousands of confirmed worlds now orbit stars beyond the Solar System: hot Jupiters, sub-Neptunes, super-Earths, frozen giants, ultra-short-period planets and a much smaller population approaching Earth's scale.
But discovering a planet and discovering another Earth are very different problems. Start with 6,354 confirmed planets, then ask whether a world is small enough to plausibly be rocky, whether it receives temperate stellar irradiation, whether its radius and mass are actually measured, whether those measurements have useful uncertainties, and whether the host star lies inside the calibrated range of the adopted habitable-zone model.
The remarkable thing is no longer how many planets we have discovered. It is how few survive a serious search for something remotely Earth-like.
This project begins there — not with the assumption that another Earth exists in the catalogue, but with the question: what does the evidence actually allow us to say?
02 — 164,209 Records. 15 Worlds Survive the First Cut.
Finding Earth 2.0 begins with the archives. The current build brings together 164,209 provenance-tracked source records, including catalogue measurements, stellar parameters, candidate records, spectroscopy metadata and supporting astronomical information.
The search then narrows: 6,354 confirmed planets → 174 nominal conservative habitable-zone worlds → 15 worlds that are both in the nominal conservative habitable zone and below the project's adopted 1.6 R⊕ first-cut radius threshold.
Those 15 are not declared "Earth 2.0". They are simply the worlds for which the available measurements justify asking the next question: how much do we actually know about them?
03 — The Most Important Number Might Be 1.
Planetary radius tells us how large a world appears. An independent mass measurement adds a second physical constraint and helps us estimate bulk density, surface gravity and escape velocity without deriving everything from the radius itself.
Among the 15 small worlds surviving the nominal conservative-habitable-zone first cut, the current analysis finds 1 with an independently measured mass. Other entries may rely on a radius-based inferred mass or an RV minimum mass, M sin i, which carries different information.
Finding planets is no longer the whole problem. Measuring Earth-sized planets well enough is.
04 — Earth-Like Is Not the Same as Habitable.
Finding Earth 2.0 deliberately separates five concepts that are often collapsed into a single headline.
Earth similarity
Do measurable bulk properties resemble Earth? Similarity is a comparison, not evidence of habitability.
Habitable-zone position
Does incident stellar flux fall within stated climate-model boundaries? Being inside a habitable zone does not prove liquid surface water exists.
Rocky plausibility
Is a predominantly rocky composition still plausible given the measured size and other constraints? Radius alone does not reveal geology or interior structure.
Atmospheric observability
Could present or future instruments realistically characterise the atmosphere? Observability is an instrumental question, not a habitability score.
Evidence for biology
No metric in this project estimates a probability of life, and no planet is presented as having established biological activity.
Five questions. Five different meanings. They should never be collapsed into one.
05 — What Would Earth Look Like to Us From 100 Parsecs Away?
Put Earth around a distant star and move it outward: ten parsecs, thirty, one hundred. The familiar image disappears. Continents, cloud systems, oceans and cities are not what a distant observer would normally receive as direct evidence.
What may remain are indirect observables: a periodic stellar dimming that constrains orbital period and radius, stellar irradiation and equilibrium-temperature estimates, perhaps a radial-velocity mass, and — in favourable cases — fragments of atmospheric information from spectroscopy.
Finding another Earth may be easier than proving that it is another Earth.
The gap between a beautiful planet rendering and what a telescope actually measures is one of the central ideas of this project.
06 — A Candidate Is Not a Discovery of Life.
A world can be potentially rocky, temperate in its received stellar flux and close to Earth's size while the most important questions remain unanswered: atmosphere, pressure, climate, surface conditions, liquid water and biology.
Finding Earth 2.0 therefore does not convert an Earth Similarity Index, a habitable-zone classification or a spectral feature into a made-up "chance of life".
Interesting is not the same as inhabited.
07 — The Missing-Mass Problem
Two planets can have similar radii and radically different compositions. One may be dense and rocky; another may contain a substantial volatile inventory or retain a gaseous envelope.
Radius asks how big? Mass begins to answer how much material? Together they constrain how dense? That is why independent mass measurements are disproportionately valuable for small, temperate candidates.
A radius tells us how big a world is. It doesn't always tell us what the world is.
08 — Every Number Has a History.
A catalogue value such as Radius: 1.07 R⊕ can look like a single fact. Behind it may sit an observing programme, stellar-radius analysis, transit-depth measurement, published uncertainty, archive ingestion, unit validation, Monte Carlo propagation and a final ranking contribution.
Finding Earth 2.0 treats that chain as part of the result. Derived values remain distinguishable from measured values, inferred masses remain distinguishable from independent measurements, and missing information remains missing instead of being quietly replaced by convenient numbers.
Don't trust the score. Trace it.
Explore the data and provenance or the methods behind the calculations.
09 — What 4,000 Possible Versions of One Planet Look Like
A catalogue usually presents one central value, but measurements are not infinitely precise. Radius, mass, stellar temperature, luminosity, orbital parameters and incident flux may all carry uncertainty, and those uncertainties propagate into derived quantities and rankings.
The current pipeline uses 4,000 Monte Carlo draws per planet. Each draw is a physically possible realisation under the reported measurement uncertainties and adopted model assumptions.
A tightly constrained candidate forms a compact posterior cloud; a poorly measured candidate spreads across a much larger region. That width is itself scientific information.
A planet isn't one number. It's a probability cloud.
10 — The Telescope Shapes the Universe We Think We Know.
The observed exoplanet catalogue is not the intrinsic Galactic planet population. Transit surveys favour particular geometries and detectable dips over finite observing baselines; radial velocity is more sensitive to systems producing measurable stellar reflex motion; microlensing and direct imaging probe different parts of parameter space again.
Geometry, sensitivity, mission duration, stellar properties, target selection and follow-up priorities all shape what enters the catalogue.
The exoplanet catalogue is not the Universe. It is the Universe filtered through our instruments.
11 — Meet the Worlds That Survived
The ranking is generated by the analysis rather than by a hand-curated favourites list. In the current build, the leading candidates include:
- Proxima Cen b — 1.3 pc · Earth-2.0 index 0.876 · msini lower limit
- GJ 1061 d — 3.7 pc · Earth-2.0 index 0.875 · measured
- GJ 1002 b — 4.8 pc · Earth-2.0 index 0.849 · msini lower limit
- Wolf 1069 b — 9.6 pc · Earth-2.0 index 0.839 · msini lower limit
- Teegarden's Star c — 3.8 pc · Earth-2.0 index 0.823 · msini lower limit
These worlds are not labelled confirmed Earth twins. A more defensible description is: worlds worth measuring better.
See the complete ranking and candidate atlas.
12 — There Is No Single Definition of the “Best” Planet.
One candidate can have high Earth similarity and habitable-zone consistency but weak observational confidence. Another can be less Earth-like while being much more precisely measured and far easier to characterise.
The project therefore exposes separate axes for Earth similarity, conservative habitability, observational confidence and characterisation potential.
The ranking is not designed to eliminate disagreement. It is designed to expose it.
13 — What Would Convince Us?
A useful ranking should reveal what is missing. For one candidate, the most valuable next step may be an independent mass; for another, improved stellar parameters, repeat spectroscopy, stellar-activity monitoring, better orbital constraints or a genuinely informative atmospheric spectrum.
The stronger question is often not "which candidate ranks first?" but which observation would reduce the most important uncertainty?
The ranking isn't the end of the search. It tells us what to measure next.
Explore the follow-up analysis.
14 — No Aliens. No Hype. Just the Evidence.
Potentially habitable exoplanets are easy to sensationalise. This project deliberately avoids invented biosphere probabilities, fake spectra, imaginary oceans presented as observations, and the automatic promotion of a habitable-zone location into a claim of habitability or life.
Instead: measurements, uncertainties, published evidence, reproducible analysis and traceable provenance.
Scientifically informed planet renders are interpretations, not direct images. Spectral bands are not automatically detections. Habitable-zone membership is not evidence of an inhabited world.
Wonder doesn't require exaggeration.
15 — Is There Another Earth?
Thousands of planets are now known. Hundreds occupy regions that can be called temperate under stated models. A much smaller group begins to resemble some of the physical conditions associated with Earth, while major gaps remain in what we know about nearly all of them.
Is there another Earth?
We don't know. That is the scientifically correct answer.
But public archives are growing, measurements are becoming more precise and atmospheric characterisation is reaching worlds that were previously inaccessible. We may not yet know where Earth 2.0 is, but we can now search the evidence at planetary scale.
Project and author
Finding Earth 2.0 in Distant Worlds is an open-source research project by Biswajit Jana. The software is MIT-licensed; source astronomical datasets remain governed by their originating archives.
Source repository · Biswajit Jana on GitHub
earth2 v1.0.1 · analysis generated 2026-08-28 20:00 UTC · Python 3.9.19 · runtime 76.1s