Photo 1 Research

I'm interested in developing tools to explore enormous galaxy datasets or surveys.

I have designed a new clustering algorithm to detect galaxy clusters, the Bayesian Cluster Finder (BCF, Ascaso et al. 2012), written in Python language. This method uses Bayesian statistics and unsupervised Machine Learning methodologies to find clustering of galaxies in n-dimensional spaces. I have successfully applied this methodology to detect galaxy clusters in many big datasets such as the Deep Lens Survey (DLS, Ascaso et al. 2014a), the CFHTLS-Archive Research Survey (Ascaso et al. 2012), and the ALHAMBRA survey (Ascaso et al. 2015a). The cluster and group catalogue obtained for the ALHAMBRA survey can be found in this link.

In paralel, I have also developed PhotReal (Ascaso et al. 2015b), a Python tool to transform existing cosmological simulations into more realistic mock catalogues of galaxies. I have applied this algorithm to several mock catalogues in the ALHAMBRA survey (Arnalte-Mur et al. 2014, Ascaso et al. 2015a), the J-PAS Survey (Zandivarez et al. 2014, Ascaso et al. 2016) and the Euclid Survey (Ascaso et al. 2015b). The mock catalogues for the Euclid and LSST surveys are publicly available in this link.

Complementarily, I have been analysing the previous simulation galaxy datasets to create regression models by using creative feature engineering to calibrate scaling relations in clusters (Ascaso et al 2016, 2017).

Using these results, I am building inference models of the counts of galaxy clusters to make predictions on the main composition of the Universe. To do this, I am using techniques such as bootstrapping and Markov-Chain-Monte-Carlo (MCMC), mostly in Python and C++ environments.

Furthermore, I have explored the change in the statistical properties of the Brightest Cluster Galaxies as a function of their distance (Ascaso et al. 2011, 2014b). To do this, I used Python and IDL to perform image manipulation and data visualization and analysis, and inference and statistical packages to evaluate the confidence of our results.

Begoña Ascaso

Begoña Ascaso

Université Paris Diderot
Laboratoire Astroparticule et Cosmologie (APC)

10, rue Alice Domon et Léonie Duquet
75013, Paris (France)

Skype: laurentinaSL
E-mail: ascaso [AT]