Which candidate Habitable Worlds Observatory (HWO) direct-imaging targets have complete pre-launch observational coverage -- stellar parameters, chromospheric activity, multiplicity (WDS), and HGCA astrometric acceleration -- and for which stars is a significant HGCA acceleration not matched by an archive-listed planet?
This project cross-matches the NASA Exoplanet Archive's 164-target Direct
Imaging ExEP Mission Star List (Mamajek & Stapelfeldt 2024) against the
Hipparcos-Gaia Catalog of Accelerations (Brandt 2021) and reports, per
star and per category, whether a real measurement exists. Two archive
fields (known-planet flag, circumstellar-disk flag) are tracked separately
and deliberately excluded from the completeness score: they are
populated for essentially every row in this list regardless of whether any
actual companion or exozodi observation was ever made, so including them
would silently inflate every star's score by a fixed, uninformative amount
-- see docs/LIMITATIONS.md. This project does not rank targets by a
single combined score -- see docs/METHODS.md for why that distinction
matters.
git clone https://github.com/Biswajit1999/hwo-target-evidence-atlas.git
cd hwo-target-evidence-atlas
python -m pip install -e ".[dev]"
pytest -q # 23 fast tests, fixture-based, no network
hwoevidence build --out results/
python scripts/make_figures.py # regenerate results/figures/See docs/REPRODUCIBILITY.md for full detail.
Across the 164 targets (results/evidence_matrix.csv, generated by
hwoevidence build, one row per target -- target_id is the stable
per-row key; several targets share a Hipparcos number with a wide-binary
companion, so hip_name/hip_number are not unique row keys): real
observational coverage is 100% for stellar parameters, 99.4% for
chromospheric activity, 97.6% for HGCA astrometric-acceleration entries,
and only 66.5% for multiplicity (a Washington Double Star catalog
cross-match) -- the clearest, most actionable gap in this list. Mean
evidence_completeness across these four real coverage categories is
90.9%. (The known-planet-flag and circumstellar-disk-flag archive fields
are separately 100% populated for every target -- shown in gray in the
figure -- but this reflects the archive's bookkeeping, not real companion
or exozodi surveys; see docs/LIMITATIONS.md.)
31 of 164 targets (28 after Benjamini-Hochberg false-discovery-rate
correction across all 164 comparisons, hgca_significant_accel_fdr in
results/evidence_matrix.csv) show a statistically significant (p < 0.01)
HGCA astrometric acceleration alongside no archive-listed planet
(results/hgca_accelerators_without_listed_planets.csv). This is not
a catalog disagreement or a confirmed companion: sy_planets_flag == "N"
means no archive-listed planet, not a confirmed single-star determination
-- the acceleration could come from a stellar, brown-dwarf, or
planetary-mass companion the archive simply hasn't listed as a planet. See
docs/METHODS.md and docs/LIMITATIONS.md for exactly what this does and
does not establish.
- Not a claim that any specific star does or does not host a planet, companion, or exozodiacal dust beyond what the cited catalogs directly report.
- Not an official Habitable Worlds Observatory, ExEP, or NASA Exoplanet Archive product; not affiliated with or endorsed by NASA, STScI, JPL, or the ExEP program.
- Not a target-selection recommendation -- mission target selection is a decision this project has no role in.
src/hwoevidence/ catalog adapters, evidence matrix, conflict detection, ranking
tests/ fixture-based + live-network tests
docs/ METHODS, DATA_SOURCES, LIMITATIONS, REPRODUCIBILITY, VALIDATION, CLAIMS
results/ generated outputs only (see results/README.md)
See CITATION.cff. This project builds on and cites Mamajek & Stapelfeldt
(2024, ApJS 272, 32) and Brandt (2021, ApJS 254, 42).
See CONTRIBUTING.md.
See ACKNOWLEDGMENTS.md.