Sign in

Marcos Huerta

@marcoshuerta.com
603 followers 868 following 629 posts

Past: PhD in Astrophysics from Rice University -> Science policy at the US Department of Energy. 
Present: Data scientist living in Virginia. I am also on Mastodon, @marcoshuerta@vmst.io marcoshuerta.com

PostsRepliesMedia
Marcos Huerta @marcoshuerta.com · 29/09/2026
This doesn't seem … good?
Bar chart showing Anthropic negative earnings increasingly negative to -$42 billion in FY 2025.
010
Marcos Huerta @marcoshuerta.com · 29/09/2026
The python to parse all these files is such a monstrosity - which is why it's one of my few repo's that isn't public! 😭 Thankfully I now load a (slow) json API with all the vote data, but behold this mess from 3 years ago, complete with commented-out bespoke debugging print statements.
A horrible mess of python code.
110
Marcos Huerta @marcoshuerta.com · 20/07/2026
Have you seen this Muppet?
The Muppet named Dr. Teeth, ostensibly to the leader of the band Electric Mayhem - yet not present at the FIFA World Cup halftime.
111
Marcos Huerta @marcoshuerta.com · 18/07/2026
I have a way to download drive data from the Kia Owners portal for my EV6, which I use to make a few visualizations on my web site. But recently, I discovered a massive data problem: the driveTrain usage looks to be stored as a 16 bit signed integer, no drive has a value above 32,767. #dataDS
A scatter plot of miles vs EV energy used by the drivetrain. linear trend until 32767 where every value is max 32767
220
Marcos Huerta @marcoshuerta.com · 15/07/2026
Philadelphia booing Joe Carter at the 1996 All Star Game.
010
Marcos Huerta @marcoshuerta.com · 19/05/2026
I think about this 2018 tweet often:
Neel Kashkari
@neelkashkari
17 Apr 2018
There's something very emotional for firms about wages.  Paying more for oil, or steel, or corn is no big deal. That's just the market. But if they can't find workers at the wages they are used to paying, that's a historic worker shortage.
120
Marcos Huerta @marcoshuerta.com · 04/04/2026
What is not widely known is that the SLS rocket was essentially created by the Senate to keep shuttle-era contractors afloat. It uses a lot of older space shuttle technology. www.nbcnews.com/science/spac...
Saving the space shuttle economy
As NASA prepared to retire its fleet of space shuttles in 2010 and 2011, Congress authorized creating the next-generation Space Launch System rocket as a way to soften the blow of the shuttle program's conclusion. Senators in states like Florida, Alabama and Utah jockeyed to save jobs for longtime NASA partners and space shuttle contractors.
"It was birthed by Congress itself," Dreier said.
In the end, the rocket was designed to use space shuttle components (the space shuttle main engines were upgraded into the Space Launch System's RS-25 core stage engines), made in large part by the existing shuttle workforce.
100
Marcos Huerta @marcoshuerta.com · 17/01/2026
Pretty sure Wittman is gone regardless.
The National Democratic Redistricting Committee, the party’s group leading the charge on redistricting, confirmed to POLITICO it has presented two new maps to Virginia lawmakers. One remakes the map into a 9-2 configuration that only provides cover for districts held by Republican Reps. Ben Cline and Morgan Griffith from being eliminated. Another proposed map largely leaves intact Griffiths’ seat and, if approved, could deliver a 10-1 map, a development first reported by Punchbowl News.
210
Marcos Huerta @marcoshuerta.com · 18/11/2025
050
Marcos Huerta @marcoshuerta.com · 07/09/2025
Face with vampire fangs and green eyes.
020
Marcos Huerta @marcoshuerta.com · 03/08/2025
One more plot, now for all sibship sizes. The coin toss + dynamic stopping model reproduces the same ever-growing binomial deviations shown in the paper (larger at sibship size 5 than at size 3). #dataBS (I figured out how to give Plotly facets different x-axis ranges!)
Deviations from binomial, for simulated data (coin toss + dynamic stopping probability). This looks very similar to Wang et al 2025 Figure 2, including ever-larger binomial deviations as sibship size grows.
000
Marcos Huerta @marcoshuerta.com · 02/08/2025
Here is a bar graph like the paper's figure 2, but this is for my simulated data where the probability of a boy is a 51.9% coin toss & the stopping probability is reduced if the children are all boys or all girls. It looks strikingly like Figure 2 in the paper.
a bar graph shoving relative devation of .06 for N male offspring of 0 and 3.
110
Marcos Huerta @marcoshuerta.com · 02/08/2025
The chance of a third boy after two boys at sibship size 3 was 49.5%, well below the study population of 51.9%. All third children after 2 boys? 51.2%. The deviations from the binomial aren't driven by a third child clustered with the first two, but by the distribution of the first two kids.
First Two Births, sibship Size 3

Table with numbers, but FF and MM are over the binomial expectation, and FM and MF are under.
110
Marcos Huerta @marcoshuerta.com · 02/07/2025
National Polls in 2021 showed people thought restrictions were worth it; what is the evidence this was a salient issue in the 2021 VA elections?
Majorities say restrictions on activity have hurt businesses, limited people's lifestyles - but see the public health benefits as having been worth the costs % of U.S. adults who say restrictions on public activity in the U.S. due to the coronavirus outbreak have …
A LOT
69%
SOME
26%
Hurt businesses and economic activity
Kept people from living their lives the way they want
Helped prevent hospitalizations and deaths from the coronavirus
58
31
32
41
Helped to slow the spread of the coronavirus
31
41
% of U.S. adults who say the public health benefits of restrictions on public activity to slow the spread of the coronavirus ..
Have NOT been
Have been
worth the costs
worth the costs
37%
62%
Note: Respondents who gave other responses or did not give an answer are not shown.
Source: Survey conducted Aug. 23-29, 2021.
"Majority in U.S. Says Public Health Benefits of COVID-19 Restrictions Worth the Costs, Even as Large Shares Also See Downsides"
PEW RESEARCH CENTER
100
Marcos Huerta @marcoshuerta.com · 29/06/2025
I added some follow up my original post, now that the jackpot was won on Friday. Using the drawings starting Jan 21 as the baseline was the most favorable comparison I could have made. Other recent drawings under the old rules from earlier in 2024 had jackpots that grew even faster. #dataBS
Graph comparing mega millions jackpot growth for several recent start dates. The jackpot growth under the new rules of April 22 lag behind the other recent jackpots under the old rules.
020
Marcos Huerta @marcoshuerta.com · 21/06/2025
The Mega Millions jackpot is up to $326M. They changed their prizes and upped the ticket price to $5 claiming his would lead to "faster growing" jackpots. But it's actually growing slower than it did under the old rules, which was $344M under the old rules at draw 19.
Graph comparing old and new mega millions rules. New Game jackpots are ~5% lower than the previous rules after draw 7.
140
Marcos Huerta @marcoshuerta.com · 21/04/2025
I'm still wrapping my head around how much solar power is being generated in Texas. At this very moment, in Texas, solar is 45% of energy generation. But, there's a bill in the legislature (that passed the senate) to sabotage all this solar (and wind) growth. www.texastribune.org/2025/04/16/t...
Fuel mix for ERCOT (Texas) real time at 13:11 Central. Solar is 45.4%. Wind is 12.2%Graph showing large new solar installation in Texas especially starting in 2023.
11311
Marcos Huerta @marcoshuerta.com · 09/03/2025
Seems bad.
000
Marcos Huerta @marcoshuerta.com · 23/02/2025
I know you have seen this before but it is the perfect response to the magazine’s nonsense. And my favorite accidental astrology debunking ever.
Okcupid match % by zodiac-zodiac pairing. Every pairing is 60%. One is 59%.
040
Marcos Huerta @marcoshuerta.com · 16/02/2025
Bird flu even got the chickens at the kid museum.
Egg/chicken exhibit at a children’s Museum
010
Marcos Huerta @marcoshuerta.com · 21/12/2024
After floundering on #AdventOfCode for the last week, my reaction to Day 21:
George Constanza leaving on a high note. "That's it for me!"
150
Marcos Huerta @marcoshuerta.com · 28/11/2024
Thanks to the FontAwesome extension for @quarto.org I now have a Bluesky icon linking to my Bluesky profile on my website:
Navbar from web site showing Mastodon, Bluesky, Github, LinkedIn, RSS, Mail, and Search icons.
110
Marcos Huerta @marcoshuerta.com · 15/11/2024
I tried out the Blue Sky API for searching via python, dusting off my old "TwitterWordle" code that solved Wordle based on shares of 🟨⬜⬜🟩🟩, etc. It still works! It needed ~300 posts from the search_posts endpoint to solve today's puzzle. (Image has no spoilers.) #dataBS
Solving Wordle from shared posts. (Contains a bar graph of scores for the top candidate solutions)

394 tweets
995 score patterns. 90 unique.
Wordle 1245 initial signal low 1.01. Iterating for better parameters
Iterated to a better signal with min_count 2 and penalty -4.50+08 61.59%, (85/138) valid final guess patterns found. Impossible pattern count: 5.
Wordle 1245 solution hash: c20e5a2ac077a8e0b2e6d575ce60a59d9a7f01558ba2d4933b2faddef843a0a6. 2.6 STD above mean. 1.13 above runner up.
Confirm Wordle 1245 solution from https://www.nytimes.com/svc/wordle/v2/2024-11-15. json
Solution is True
250
Marcos Huerta @marcoshuerta.com · 13/11/2024
re: unix commands, I would say there is a high chance you have seen this xkcd but just in case for anyone who hasn’t: imgs.xkcd.com/comics/tar_2...
Xkcd: disarm a bomb with valid tar command. The hero can’t do it and apologizes.
080
Marcos Huerta @marcoshuerta.com · 23/10/2023
Unless you are a *fan* of any of the remaining four teams, in which case it’s one of the most stressful and gut-wrenching days of the year! 😬 Playoff baseball! 🫣
040