Sabina Sagynbayeva
@sabinastro.bsky.social
(astro)Physics PhD candidate @ Stony Brook. Astrobites alum. From Kazakhstan 🇰🇿. She/her.
Musical theater nerd, so pretty much insufferable.
Statistics, planets, and stellar cartography.
https://ssagynbayeva.github.io
Musical theater nerd, so pretty much insufferable.
Statistics, planets, and stellar cartography.
https://ssagynbayeva.github.io
Ah I just read the thread. It’s also not cool to do after the deadline passed. 🥲
November 5, 2025 at 3:42 PM
Ah I just read the thread. It’s also not cool to do after the deadline passed. 🥲
Wait what happened? I just applied there last week….
November 5, 2025 at 3:39 PM
Wait what happened? I just applied there last week….
Honestly this sounds fun even to me at my age too
October 11, 2025 at 2:58 AM
Honestly this sounds fun even to me at my age too
This breakfast looks so goodddd🤤
October 7, 2025 at 1:37 PM
This breakfast looks so goodddd🤤
Omg that was my attempt 8 months ago 🤣 the actual model code is in Zenodo: doi.org/10.5281/zeno...
But I do need to update my github repo! And I mostly used celerite but I do want to clean up and create a notebook with notes on sampling and stuff. :)
But I do need to update my github repo! And I mostly used celerite but I do want to clean up and create a notebook with notes on sampling and stuff. :)
Rotation Periods for Stars in Open Cluster NGC 6819 From Kepler IRIS Light Curves
Samples from the gyrochrone fitting with a Mixture Model from the paper Rotation Periods for Stars in Open Cluster NGC 6819 From Kepler IRIS Light Curves from Sagynbayeva, Colman, & Farr 2025.'gyrochr...
doi.org
October 7, 2025 at 1:14 AM
Omg that was my attempt 8 months ago 🤣 the actual model code is in Zenodo: doi.org/10.5281/zeno...
But I do need to update my github repo! And I mostly used celerite but I do want to clean up and create a notebook with notes on sampling and stuff. :)
But I do need to update my github repo! And I mostly used celerite but I do want to clean up and create a notebook with notes on sampling and stuff. :)
Will never unsee it
October 6, 2025 at 6:49 PM
Will never unsee it
Anyways, like, comment, and hire me PLEASE!!!
October 3, 2025 at 1:32 PM
Anyways, like, comment, and hire me PLEASE!!!
In the end, we have a beautiful gyrochrone that you should take with a grain of salt just like any other gyrochronal sequence! Also my sample is the largest sample of rotation periods in an open cluster to date! Yay!
October 3, 2025 at 1:32 PM
In the end, we have a beautiful gyrochrone that you should take with a grain of salt just like any other gyrochronal sequence! Also my sample is the largest sample of rotation periods in an open cluster to date! Yay!
So if you assume that the uncertainties in period in period and temperature are gaussian, you can imagine that the point is sliding in those error bars until it decided whether to fall on our 5th order Legendre polynomial or not.
October 3, 2025 at 1:32 PM
So if you assume that the uncertainties in period in period and temperature are gaussian, you can imagine that the point is sliding in those error bars until it decided whether to fall on our 5th order Legendre polynomial or not.
So the mixture model took the probability weights, and assigned some stars to the background based on their location in the period-temperature space. But we accounted for uncertainties in period and temperature too!
October 3, 2025 at 1:32 PM
So the mixture model took the probability weights, and assigned some stars to the background based on their location in the period-temperature space. But we accounted for uncertainties in period and temperature too!
Instead, I wrote a Bayesian mixture model. The premise of the model is that some of the stars lie on the gyrochrone, but some of the are what we called “background stars”.
October 3, 2025 at 1:32 PM
Instead, I wrote a Bayesian mixture model. The premise of the model is that some of the stars lie on the gyrochrone, but some of the are what we called “background stars”.
For gyrochrone, I wanted to just fit a quick polynomial but Will said he’d remove his name from the paper if I did that. 😂
October 3, 2025 at 1:32 PM
For gyrochrone, I wanted to just fit a quick polynomial but Will said he’d remove his name from the paper if I did that. 😂
We had to make some quality cuts, remove giants, and we were left with 271 reliable rotations.
October 3, 2025 at 1:32 PM
We had to make some quality cuts, remove giants, and we were left with 271 reliable rotations.
So to account for that, my GP kernel was a sum of the Rotation term (i.e. the period) and the term that you can think of as a RealTerm in celerite so that it models the red noise.
October 3, 2025 at 1:32 PM
So to account for that, my GP kernel was a sum of the Rotation term (i.e. the period) and the term that you can think of as a RealTerm in celerite so that it models the red noise.
I used gaussian processes to create a Bayesian model to extract periods from pretty low-amplitude light curves. We looked at PSDs and they still had this exponential decline that was the “red noise” in the data.
October 3, 2025 at 1:32 PM
I used gaussian processes to create a Bayesian model to extract periods from pretty low-amplitude light curves. We looked at PSDs and they still had this exponential decline that was the “red noise” in the data.
Getting periods from the crowded fields is HARD! So Isabel had to do her own photometry with superstamps, Will came up with a new-ish de-trending method, and I was left with maximizing likelihoods! 😄
October 3, 2025 at 1:32 PM
Getting periods from the crowded fields is HARD! So Isabel had to do her own photometry with superstamps, Will came up with a new-ish de-trending method, and I was left with maximizing likelihoods! 😄