Eamonn Kerins
banner
eamonnkerins.bsky.social
Eamonn Kerins
@eamonnkerins.bsky.social
160 followers 130 following 81 posts
Jodrell Bank astrophysicist and exoplanet scientist. Swimmer, cyclist, runner. Sometimes all three consecutively. Occasional BBC TV space pundit. Prone to career-threatening scicomm and fundraising activities. Own views. https://linktr.ee/eamonn_kerins
Posts Media Videos Starter Packs
Proud to be part of this amazing team.
The Roman Galactic Exoplanet Survey (RGES) #PIT is building tools to help reveal new worlds across the Milky Way. 🌍✨

From microlensing to planetary demographics, they’re paving the way to support the exploration of a hidden population of exoplanets with #NASARoman.

🔗 rges-pit.org
Can't tell for sure, to be fair. But i know this one is official: @nancyromansci.bsky.social
Attending (virtually) a @nasaromantelescope.bsky.social collaboration meeting this week where we are getting our heads around the science we'll do with 100,000 new exoplanets over its five year nominal mission. For perspective, the last 30 years has delivered 6000. Roman will be mind blowing! 🔭🧪🪐
Roman is expected to find 60,000-200,000 transiting planets, alongside mass measurements of cooler planets from microlensing.
If you plot only the discovery technique measurement, there's only place on the plot for each planet. Symbol shapes indicate if radius ratio (circle) or mass ratio (triangle) so there's no ambiguity. No need to convert in my view. I don't mind apples and pears both occupying the same fruit bowl.
And, to clarify, I'm plotting the discovery technique measurement. So one unique location per planet. Demographic studies are normally based on measurements of the discovery technique.
And these quantities are also closer to what is often directly measured. Eg transit lightcurves measure radius ratio, not radius. Microlensing measures mass ratio, not mass.
Sure, but the usual plots exclude one or more techniques by choosing either mass or radius as one of the axes. This plot doesn't choose, so can represent all. Also mass ratio, rather than mass, appears to be more fundamental. So, plotting both mass and radius by host ratio makes sense.
I like the way that the four #exoplanet detection methods are relatively well separated when plotting their bulk properties as ratios relative to their host. The scatter shape also reminds me of a hovering #hummingbird. So, I'm going to call this plot the hummingbird plot. 🧪🔭🪐
Been bad mouthing #AI engines recently. So, for balance, i got #GPT4 to do some useful #exoplanet stuff today. From a csv i downloaded from the #NASA #ExoplanetArchive it made a clickable version of the plot below that takes you to the info page on that planet. Damn useful! 🧪🔭🪐
Been bad mouthing #AI engines recently. So, for balance, i got #GPT4 to do some useful #exoplanet stuff today. From a csv i downloaded from the #NASA #ExoplanetArchive it made a clickable version of the plot below that takes you to the info page on that planet. Damn useful! 🧪🔭🪐
I guess some are better than others - Claude Opus 4.1 is not so easily fooled, it seems. Not on this test anyway.
Something else #GPT5 is really bad at is updating the content of lecture slides on #exoplanets. I thought it might actually do a reasonable job on this type of task, so i fed it one lecture of content. Nope, it was awful! It either made no changes whatsoever, or it completely ruined everything.
So, #GPT5 is supposed to have "PhD level" intelligence, according to the hype. Every time a new #AI chatbot is released, I test it with a fictitious situation in which I attempt to con a friend out of £5. A ten year old could easily spot the con. GPT5, like its predecessors, cannot.
So, #GPT5 is supposed to have "PhD level" intelligence, according to the hype. Every time a new #AI chatbot is released, I test it with a fictitious situation in which I attempt to con a friend out of £5. A ten year old could easily spot the con. GPT5, like its predecessors, cannot.
And here's what it looked like - magical!
I mentioned on #BBCBreakfast yesterday that I would see in the #SturgeonMoon with a moonlight dip. Being in water is a wonderful vantage point to observe the sky.
#nightswimming #fullmoon
If you're in the UK and hoping to catch the #PerseidMeteorShower over the next few days, here's some pointers from my chat on #BBCBreakfast this morning with Charlie and Naga.
My daughter did this sketch a few years ago. Seems an appropriate evening to give it an airing.
#OasisLive25
Reposted by Eamonn Kerins
The Met Office is currently looking for people to suggest up to 5 names for storms...

It would be great if thousands all suggested the same 5 names (in alphabetical order as that's how storm names appear):
BigOil, BP, Equinor, Exxon, Shell.

www.metoffice.gov.uk/forms/name-o...
Name our Storms
www.metoffice.gov.uk
Whilst Top Gear is not on our screens, I'm think there's a gap for Sky At Night to have its own version of #TheStig. What do you think @chrislintott.bsky.social ? 👽
Beautiful swim at Combs Reservoir near Buxton this morning. But the large sandy looking shore line area certainly testifies to the effects of the recent dry weather spell.
Another orbit complete, and a lovely card from my daughter, showing some family moments over the last 23 orbits.

I'm not tearing up, honest, I just sweat a lot...