My philosophy is that projects have a beginning, a middle and an end. In that order. So research projects are pretty strickly organized if I can help it with a clear end date in mind as well as an organization in cloud storage (dropbox, overleaf) to avoid the hunt-that-file again for me and the student. 

I have a drawer of outstanding astronomy problems to pick from and I try to work out a problem that hits on a student's interests. 

I am happy to work with anyone under one condition: we both take our investment (time) seriously. If I dedicated time to your research, I expect to hear from you on that date and time. It also means that I consider any researcher's time worth something: references, money or credit and I will work with what works for you. 

Undergraduate research should be fairly short (a few months) so I will not encumber you with data-reduction or the like. Get one or more catalogs, make a plot answering an outstanding question based on that catalog and put your science into words (spoken and written down) are the key elements of research projects with me. 

Other than that, you can look at the research page to give you an idea on what I work on and we can chat and design a project. 

Accidentally overlaps of two galaxies allows us to study the foreground galaxy's optical transmission. We use the argument that both galaxies should be rotationally symmetric to measure and map the missing light of the background galaxy. 

With citizen science, we now know of a lot of new overlapping galaxies and several have been observed with the Hubble Space Telescope.

Where do galaxies have the most dust? Is it all in dramatic dust lanes or is there more diffuse parts? How far out do galaxies show dust? How much is there? 

There are multiple projects we can do, ranging from single objects to small surveys. I am working on an analysis package that allows for this type of analysis in Python environment. 

One pair (UGC 3995), I have a nice Integral Field Unit (IFU) observation of. This is more complex data (every pixel has a spectrum as well) but it could be a very good experience for those who want to go in depth into astronomical data analysis. 

The Looking At the Distant Universe with the MeerKAT array has one ultradiffuse galaxy in it. How these faint, fuzzy and mostly old galaxies came to look this way may have something to do with their gas supply. This project is to examine all the information we have on this galaxy and place it into context of other ultra-diffuse galaxies known. Has this galaxy a ready supply of gas that could be turned into stars or has it run out? How does that compare to other UDG populations? Is that typical?

The Galaxy And Mass Assembly (GAMA) survey has a wealth of data in catalogs. As part of this collaboratation, we would examine how the history of forming stars depends on the number of spiral arms galaxies have. A previous student project found that galaxies with more arms grew slower relatively speaking. Has this always been the case? 

There are catalogs of the total angular momentum of nearby galaxies. Taking the orbits of stars and gas in galaxies, how does it depend on galaxy morphology? There is evidence the presence of a central bulge is important but what about a central bar? Combining the angular momentum catalog with the S4G catalog of stellar morphology, we can plot the effect of bars on the mass-total angular momentum relation. 

This is a more substantial project where we would produce a catalog of the S4G sample with morphometrics (parameters that describe the appearance of the galaxy) for the full 2500 sample. Combined with visual classifications, the project is to see how well the morphometric space can distinguish between different visual types using simple machine learning algorithms (e.g. K-means clustering or support vector machines etc.)

This is also a more substantial project. The idea is that we would use Google colab or the dedicated ML machine to classify GAMA galaxies. There are several characteristics of galaxies that are hard to get unless a lot of telescope time is dedicated to it. So if a CNN could predict this value from an image (much cheaper) then this would be very useful. Characteristics like metallicity (enrichment with everything but Hydrogen/Helium), the presence of an active galactic nucleus (supermassive black hole) or an accurate count of the number of spiral arms are things a CNN could potentially do. The GAMA date would be used for training/evaluation in anticipation of the Rubin Observatory start of operations. 

The project is heavy on machine learning techniques and will need some programming to be completed. 

Some galaxies form stars pretty steadily, while smaller ones seem to undergo bursts of growth. By combining a short timescale and longer timescale estimate of star-formation of the galaxies in GAMA, we can explore how "bursty" each galaxy is. Are there steady star forming small galaxies, bursting big galaxies? 

This project will use the public GAMA data on H-alpha emission line (short timescale of 10s of Myr) and the ultraviolet / SED fit star-formation rate (150 Myr).

We will use the power spectrum of MHONGOOSE galaxies observed with MeerKAT in HI to see if the gas disk is dominated by 2D or 3D turbulence. Power spectra are a powerful technique to characterize the different scales of structure in an image. 

Some of the nearest dwarf galaxies are very whispy and hard to define. In this case and possible a few others, we can use the color-magnitude diagram of the individual stars in Hubble observations to map out the outline of this galaxy using a technique known as Voronoi Tesselation. We will start from the Hubble Legacy Archive catalog, select Red Giant Branch and bright Main Sequence stars to map out the old and young population of stars of galaxies as shown to the left here. 

Dust is a by-product of stars forming and perishing. The heavier atoms produced in stars are spread throughout galaxies and clump into "dust", grains of silicate or carbon with ices that modify starlight traveling through galaxies. The total amount of dust in galaxies is therefore an interesting component of their evolution and appearance. But dust is destroyed in star-formation and by harsh radiation as well. 

This project uses predictions from SIMBA simulations for the total dust masses of galaxies at different redshifts (lookback times, so earlier epochs in the universe) with observations from GAMA/G10/HST3D where dust mass was one of the results of fits to the data.

Which galaxies have (relatively) the most dust at each epoch? Do galaxies become dustier? How do the predictions and observations compare?