GALAXY FORMATION, A DATA-DRIVEN APPROACH

What is the origin of the wide variety of galaxies we observe in the Universe? We believe the answer to this question involves a combination of primordial conditions, cosmic environment (on a range of scales) and internal physical processes involving gravitational and barionic processes. Solving this cosmic puzzle is one of the greatest challenges cosmologist face today. Paradoxically, while state-of-the-art computer simulations can reproduce to a fairly good degree many properties of galaxies, we still have a poor understanding of the fundamental processes that drive galaxy formation and evolution.

Solving galaxy formation is a multi-faceted problem requiring the combined efforts of observers, theoreticians and recently, numerical experimentalist. I fall in the former category, devising cosmological numerical experiments in order to understand the fundamental physical processes shaping a galaxy's properties as well as developing new techniques to extract relevant information from the vast sea of data coming from observations and simulations.

COSMIC WEB AUTOMATED CLASSIFICATION

Galaxies form and evolve in a cosmic stage know as the "cosmic web". Understanding the elements that form the Cosmic Web is not a simple task and requires the use of techniques borrowed from areas as diverse as medical imaging, mathematical morphology and computer vision. Since my PhD I have developed two independent methods for characterizing the Cosmic Web. One based on local variations of the density field (geometry) and another on the topology of the density field.

A MULTI-VERSE AS A COSMOLOGICAL TOOL

The use of computer N-body simulations give us a great advantage over observations when studying complex physical processes by allowing full access to the variables of the system as well as their time evolution. However, conventional simulations also share fundamental limitations with observations, in particular finite halo sampling. Galaxies can be considered as poisson biased samplings of the underlying density field. The sampling density is given by the halo mass function. Regardless of the simulation's mass resolution the halo density (for a given halo mass) remains the same. This seriously limits the study of low-density regions such as voids where often there are not more than a couple of haloes with masses large enough to host luminous galaxies. However, the sampling limitation applies to a single realization. What if we could run "parallel universes" and stack them together?

THE SOCIAL MULTI-VERSE

What if we could ask "what if" questions to a simulation? For instance, what if the proto-galaxy that gave origin to the Milky Way was located inside a void instead of a wall? The MIP project uses the idea of a multi-verse to explore a large set of possibilities for a large halo population. Any halo in the simulation is replicated across the ensemble but sampling different cosmic environments. This provides a unique way of studying the relation between galaxies and their environment. A N-body simulation becomes a controlled experiment.

THE STRUCTURE OF VOIDS

Voids have been traditionally considered as simple structures, in fact just "holes" in the galaxy distributions. It turns out, voids are surprisingly complex objects. They contain an internal hierarchy of tenuous structures that can be seen in both the density and velocity fields. Understanding the structure and dynamics of such pristine environments and how this influences the galaxies they contain may hold the key to galaxy formation.

THE MILKY WAY'S COSMIC ENVIRONMENT

The fact that Edwin Hubble was able to measure the recession velocity of nearby galaxies and infer the expansion of the Universe is a feat of cosmic proportions. Unknown to him, the low velocity dispersions around the Milky Way needed to measure the "Hubble law" are a rare phenomena. Theoretical expectations and measurements elsewhere in the Universe give velocity dispersions one order of magnitude larger that the ones measured locally. This "cold Hubble flow" problem represented a mistery for more than 5 decades. The explanations ranged from dark energy to obscure effects of barions. The solution we proposed was the particular geometry and dynamics around our Galaxy. These two aspects of our cosmic environment are fully sufficient to explain the observed cold Hubble flow and even

LSS-HALO CORRELATIONS

One of the clearest examples of the LSS-galaxy connexion is the alignment between spin and shape of galaxies and their host LSS structure.

REPRESENTATIVE PAPERS

Galaxies and the Cosmic Web

Structure of the Cosmic Web

Methods