A curated collection of Jupyter notebooks introducing and applying statistical methods and cosmological inference tools to problems in cosmology.

This repository combines:

• Practical statistical tutorials for students and researchers.
• Step-by-step Cobaya tutorials for Bayesian cosmological parameter estimation and model comparison, now included as a subfolder of the main tutorials.

• Probability & Statistics Refresher – Basics of probability distributions, likelihoods, and Bayes’ theorem.
• Bayesian Inference in Cosmology – MCMC, nested sampling, and posterior exploration with toy cosmological models.
• Fisher Forecasting – Deriving Fisher matrices and forecasting cosmological parameter constraints.
• Likelihoods for Cosmology – Building Gaussian and non-Gaussian likelihoods connected to cosmological data vectors.
• Applications – Toy SN & LSS parameter estimation, combining multiple probes, impact of priors and degeneracies.

• Cobaya – Sampling multi-variate Gaussian and Ring likelihoods; running simple cosmology chains & using the get_model() wrapper in Cobaya to access all internal calculations without modifying the source code.

git clone https://github.com/your-username/Stats4Cosmology.git
cd Stats4Cosmology

jupyter lab

Open any notebook under the tutorials/ folder to explore Stats4Cosmology, or navigate to tutorials/software/ for software specific exercises.

You can run Cobaya notebooks on Google Colab. For Cobaya specific notebooks you will need:

!pip install cobaya
!cobaya-install cosmo -p .

The notebooks rely on standard scientific Python libraries widely used in astronomy and cosmology:

• numpy – numerical computing
• scipy – scientific routines
• astropy – astronomy & cosmology utilities
• matplotlib – plotting
• seaborn – statistical visualization
• jupyter – interactive notebooks
• cobaya – cosmological Bayesian analysis (for Cobaya tutorials)

• Students in astrophysics, physics, or statistics wanting hands-on cosmological inference training.
• Researchers looking for ready-to-use templates for teaching, lectures, or experimentation.
• Educators needing practical examples for cosmology and data science courses.
• Anyone interested in learning how to use Cobaya for Bayesian cosmological analysis.

Developed and maintained by Dr. Guadalupe Cañas-Herrera, inspired by real-world cosmological pipelines in the era of Euclid, probably biased by Planck 18 analysis experience.

This project is licensed under the MIT License – see the LICENSE file for details.

If you find these notebooks useful in your research or teaching, please cite:

@misc{stats4cosmology,
  author       = {Cañas Herrera, Guadalupe},
  title        = {Stats4Cosmology: Jupyter notebooks for statistical methods in cosmology, including Cobaya tutorials},
  year         = {2025},
  howpublished = {\url{https://github.com/gcanasherrera/Stats4Cosmology}}
}