Planetary Linguistics

Since fall 2018, my advisor David Kipping and I have been collaborating with Michael Collins of the Columbia University Data Science Institute to apply techniques from natural language processing to the problem of understanding multi-planet systems.

The first results of this collaboration show that a single-population multiplicity model---specifically, a Zipfian distribution---matches the observed multiplicities of Kepler planetary systems as well as dichotomous models. (Above: A comparison of "observed" multiplicities in a population of simulated planets (green bars) to the actual Kepler detections (black dots).) See our paper for further details.

We're currently working on a way to learn the "grammatical rules" of Kepler planets and classify them into categories analogous to parts of speech. Stay tuned!