I am hoping that someone will be able to point me in the right direction.  I am a graduate student at the University of Louisville, and I am looking to do a data mining and/or machine learning project involving astronomical data for my thesis project.  I will tentatively be completing my thesis in the spring of 2020, so it would be a moderately long-term project on my part.  I desire to make a meaningful and unique scientific contribution, and not to re-do analyses and determinations which have already been done.

My preference would be to collaborate with an astronomer or data scientist, but I am more than willing to work on a project without external support.  There are enormous amounts of data available; I just need a nudge into a good direction.  For instance, it would definitely be interesting to data mine the Kepler data and construct machine learning algorithms in an effort to enhance the detection of transiting exoplanets, but I fully expect that has been done already to a very high degree, and I would not expect myself to be able to improve upon it.  I am wondering whether SDSS (Sloan Digital Sky Survey) data might be a good starting point...

Any advice or input would be very much appreciated!

- Jonathan