exoplanet-parameter-estimation-model

A showcase-ready radial-velocity fitting project for estimating single-planet orbital parameters from Doppler observations.

What it does

This rewrite replaces the original script-based prototype with a small Python package that:

• models stellar radial velocity with a Keplerian single-planet curve
• initializes the period search with a Lomb-Scargle periodogram
• fits orbital parameters with differential evolution followed by local least-squares polish
• derives an approximate planet mass and semi-major axis from the recovered fit
• renders a two-panel figure with the RV fit and a velocity-Verlet orbit visualization
• can export a GIF of the best-so-far model evolving across optimizer generations
• generates reproducible synthetic datasets so the project runs without external CSVs
• ships four public NASA Exoplanet Archive radial-velocity series for real-data demos

The fitting flow is:

1. generate or load a radial-velocity time series
2. estimate a dominant period (for public archive demos)
3. evaluate a Keplerian single-planet model
4. search globally with scipy.optimize.differential_evolution
5. polish locally with scipy.optimize.least_squares
6. derive a representative planet mass and orbital radius for plotting

Quickstart

uv sync
MPLBACKEND=Agg uv run python examples/fit_synthetic.py --save outputs/rv_fit_preview.png
MPLBACKEND=Agg uv run python examples/export_fit_evolution_gif.py \
  --target 51_peg --save docs/51_peg_fit_evolution.gif

Those commands write:

• outputs/rv_fit_preview.png
• outputs/synthetic_rv_dataset.csv
• docs/51_peg_fit_evolution.gif

Public NASA observation data

The observational RV files live under data/nasa_exoplanet_archive/:

Key	Host	Literature source
51_peg	51 Peg	Marcy et al. 1997
hd_209458	HD 209458	Butler et al. 2006
70_vir	70 Vir	Butler et al. 2006
hd_3651	HD 3651	Butler et al. 2006

These are public NASA Exoplanet Archive RVC products. Provenance (archive URLs, bibcodes, instruments, stellar masses) is recorded in data/nasa_exoplanet_archive/catalog.json.

Fit any series with:

MPLBACKEND=Agg uv run python examples/fit_public_star.py --target 51_peg \
  --save outputs/51_peg_fit.png

Using real data

The project also exposes a CSV loader in src/exoplanet_est/data.py. Supported inputs can be either:

• named columns such as time_days / time_jd, radial_velocity_ms, and uncertainty_ms
• unnamed numeric columns ordered as time, value, optional uncertainty

If your file stores Doppler shift instead of velocity, pass value_kind="doppler_shift" to load_radial_velocity_csv(...).

For the packaged public archive series, use load_public_target_dataset(...).

Package layout

• src/exoplanet_est/keplerian.py - Kepler solver and RV curve
• src/exoplanet_est/optimize.py - periodogram bounds, global + local fitting pipeline
• src/exoplanet_est/nbody.py - velocity Verlet orbit integrator
• src/exoplanet_est/data.py - CSV loader, public-archive catalog helpers, synthetic generation
• src/exoplanet_est/plot.py - dark/print-themed figure creation and CSV export
• examples/fit_synthetic.py - end-to-end showcase demo
• examples/fit_public_star.py - public NASA RVC fitting example
• examples/export_paper_figures.py - regenerate paper CSVs / print PDFs
• examples/export_fit_evolution_gif.py - animate best-so-far RV/orbit frames over DE generations
• docs/exoplanet_showcase_report.tex - companion LaTeX write-up
• docs/figure_macros.tex - PGFPlots macros used by the paper
• docs/51_peg_fit_evolution.gif - showcase animation used on the README / project card

Companion report

A scientific write-up is included at docs/exoplanet_showcase_report.pdf (source: docs/exoplanet_showcase_report.tex, bibliography: docs/references.bib). Citations are numbered in order of appearance ([1], [2], …), which is convenient for an arXiv preprint. The paper presents the Keplerian RV model, fitting method, synthetic validation, and observational results on named public NASA archive series.

Paper figures are native PGFPlots plots driven by CSV tables under docs/figures/plotdata/. Regenerate everything with:

MPLBACKEND=Agg uv run python examples/export_paper_figures.py
latexmk -pdf -cd docs/exoplanet_showcase_report.tex

That export also writes light-theme vector PDFs in docs/figures/. Use --theme dark or --theme print on the fit examples when saving a single figure.

Acknowledgment

This research has made use of the NASA Exoplanet Archive, which is operated by the California Institute of Technology, under contract with the National Aeronautics and Space Administration under the Exoplanet Exploration Program.

Notes

The synthetic demo assumes an edge-on system (sin(i)=1) so the fitted semi-amplitude can be converted into a single representative planet mass for visualization.