Research overview
Most of my research focuses on cosmology, the study of the universe at the largest scales (longest distances) we can measure. I'm an observer, which means I make measurements using real data, usually from large surveys like the Sloan Digital Sky Survey, Hyper SuprimeCam, and upcoming surveys with the Large Synoptic Survey Telescope and WFIRST.
I mainly work on large-scale structure probes. These are measurements that look at the distribution of galaxies or matter in the universe in so-called "recent times". That's the last few billion years, when the universe looks roughly like it does today (with lots of galaxies everywhere and lots of space in between them). My main focus is weak gravitational lensing, though I also work with galaxy clusters.
I'm also interested in some of the technical aspects of how we make those measurements. I work on systematic uncertainty testing and mitigation: finding small problems in our measurements (because they're hard!) and fixing the problems if we can or describing them well if we can't. I'm also interested in using big data and machine learning techniques to improve the ways we make measurements.
Weak gravitational lensing is the effect in which light gets bent by gravity. As light passes through the universe, it goes past galaxies and other overdensities of matter, and it also goes past places where the density is less than average (voids). This slightly warps the galaxy images we see: it changes their shapes, makes them bigger or smaller, and changes their positions and, as a result, how far apart they seem. Because the distortions are due to gravity, these distant galaxies act light a 'backlight' illuminating the gravitational fields in the universe. These distortions to the galaxy images are very small and subtle, but if we do statistics with many distant galaxies (tens of millions or more), we can make very precise measurements. And because we're measuring the direct influence of gravity, we can use weak lensing to learn about dark matter as well as the ordinary matter like stars and planets we see around us.
Galaxies can be stuck together with gravity into clusters of galaxies (like the stars around us are stuck together into a galaxy). Galaxy clusters are the largest things in the universe bound together with gravity, so learning about them can tell us interesting things about how matter clumps together to form structures. Understanding how the universe forms structures is one of the key questions we care about in cosmology, and galaxy clusters help us figure out the answers.
This one is kind of technical, but I find it really interesting! Extracting information from pictures of the sky is hard. The light has passed through a telescope and maybe the atmosphere, which blurs things out; we take the images using digital cameras, which are very very good but not perfect; we have to make some simplifying assumptions in order to figure out how our measurements line up with our theories; and then we often combine several of those measurements, which depend in complicated ways on all those little difficulties. I spend some of my time figuring out how things can go wrong and how we can detect that so we can have high confidence in our final results.
Cosmological surveys have a lot of data: from hundreds of thousands up to billions of galaxies, with lots of rich information about each galaxy (such as its size, color, and shape). These data sets are a great target for what are called 'machine learning' techniques, which are ways of describing what the data looks like that's more complicated than what we can write down as an equation. Some of my projects involve taking machine learning techniques and applying them to data from cosmological surveys to learn about galaxies and about the size of systematic uncertainties, as mentioned above.