ShapeFit robustness for future galaxy surveys

This master's thesis focuses on the evaluation, refinement, and optimization of a pipeline that can be used for the ShapeFit approach for ongoing and upcoming galaxy surveys like DESI. One of the key strengths of ShapeFit lies in its ability to effectively extract both early-time and late-time cosmological information embedded within the matter power spectrum, offering a valuable tool for probing the universe's large-scale structure. ShapeFit captures most of the information obtained from BAO fit, RSD fit and full-modelling, by measuring the slope of the baryonic suppression feature in a more model agnostic way compared to full modelling. The primary objective of this thesis project is to evaluate the robustness and applicability of ShapeFit to ongoing and upcoming galaxy surveys, with a particular focus on the Dark Energy Spectroscopic Instrument (DESI) survey. To achieve this objective, the initial phase of the project involved a comprehensive review and documentation of including different smoothing algorithms implemented from the literature and different derivative methods which can be used within the framework of ShapeFit. This documentation serves to enhance code transparency, facilitate collaborative development efforts, and ensure reproducibility and reliability of results. Subsequent stages of the project have been dedicated to refining and optimizing the pipeline to address various numerical and physical limitations encountered during data analysis. For instance, efforts have been directed towards enhancing the modularity and flexibility of the smoothing algorithms, as well as mitigating the deviation from expected sigmoid behavior of the ratio between the matter power spectrum and the fiducial matter power spectrum caused by the proximity of the first BAO peak to the peak of the matter power spectrum. These refinements aim to improve the accuracy and reliability of cosmological parameter estimation, particularly in the context of ongoing and future galaxy surveys characterized by increasingly large and complex datasets. The last stage of the project involves comparative analyses of results obtained using different algorithms within the pipeline, aimed at quantifying the systematic error budget and identifying areas for further refinement. It encompass more extensive sensitivity analyses, including quantification of biases of a chosen assumption, exploration of the impact of varying cosmological parameters such as the primordial tilt of the power spectrum (ns), and identifying cases where the model fails, solving them when possible and including the cases where the model will not fully work for future users. Ultimately, the improved ShapeFit codebase and associated documentation will be made available to the DESI collaboration, thereby contributing to the advancement of analytical tools for galaxy surveys and facilitating robust cosmological analyses in the era of precision cosmology.