iiinova — where your digital self keeps evolving

Reposted by Nelson Tang

juanitorduz @juanitorduz.bsky.social · 23/11/2025

Here is the recording of my talk PyData Berlin 2025: Introduction to Stochastic Variational Inference with NumPyro Notebook: juanitorduz.github.io/intro_svi/ youtu.be/wG0no-mUMf0?... #pydata #berlin #bayes

youtu.be

Scaling Probabilistic Models with Variational Inference

YouTube video by PyData

0124

Reposted by Nelson Tang

Sung Kim @sungkim.bsky.social · 19/10/2025

A new 246-pages book on MCMC. "Finite Markov chains and Monte-Carlo Methods: An Undergraduate Introduction" This is a free textbook suitable for a one-semester course on Markov chains, covering basics of finite-state chains, many classical models, asymptotic behavior and mixing times,

1373

Reposted by Nelson Tang

Chelsea Parlett @chelseaparlett.bsky.social · 29/08/2025

I’m especially proud of this article I wrote about Gaussian Processes for the Recast blog! 🥳 GPs are super interesting, but it’s not easy to wrap your head around them at first 🤔 This is a medium level (more intuition than math) introduction to GPs for time series. getrecast.com/gaussian-pro...

A Gaussian process showing that the allowed time series are forced to be compatible with data

28023

Reposted by Nelson Tang

Rob Hyndman @robjhyndman.com · 11/04/2025

A new Python edition of "Forecasting: Principles and Practice" is now available online at otexts.com/fpppy/. Thanks to @azulgarza.bsky.social, Cristian Challu, Max Mergenthaler, Kin Olivares & Nixtla for making this happen. #forecasting #python

otexts.com

Forecasting: Principles and Practice, the Pythonic Way

38024

Reposted by Nelson Tang

Heather Urry @heatherurry.io · 27/03/2025

Need to explain (or understand) linear mixed effects regressions, random intercepts, and random slopes? Look no further than "A Visual Introduction to Hierarchical Models" by Michael Freeman, 2017. It's a banger!

mfviz.com

A Visual Introduction to Hierarchical Models

A visual explanation of multi-level modeling

99430

Reposted by Nelson Tang

Carl T. Bergstrom @carlbergstrom.com · 25/03/2025

With all the bad shit going on, I'm trying to spend more time reading science and math and less doom-scrolling. Today's diversion was an exploration of the James-Stein estimator, wherein we can get a better joint estimate of the means of three variables than by taking the mean of each variable.

researchgate.net

(PDF) Stein's Paradox in Statistics

PDF | On May 1, 1977, Bradley Efron and others published Stein's Paradox in Statistics | Find, read and cite all the research you need on ResearchGate

917818

Reposted by Nelson Tang

Sung Kim @sungkim.bsky.social · 25/03/2025

This has been updated to v2. arxiv.org/abs/2412.052...

0365

Nelson Tang @nelsontang.com · 20/03/2025

Curious about this too, I don’t see much here but might be because of chronological feed or I’m just not following the right people

020

Nelson Tang @nelsontang.com · 17/03/2025

That entry level job market is going to be sobering

190

Nelson Tang @nelsontang.com · 19/02/2025

Couldn’t agree more - only poor options to be found here unfortunately. Hoping improvements are coming, but until then it’s a lot of long breaks away from the phone

000

Nelson Tang @nelsontang.com · 19/02/2025

Have you considered the OnlyPosts feed?

100

Nelson Tang @nelsontang.com · 09/02/2025

Yeah I think pyenv did that for me when I used it, it lets you pick a python version and set a global default (or manually activate a different one). If you haven’t nuked your computer yet, you can at least use ‘uv python list’ to see what’s installed and do some cleanups

000

Nelson Tang @nelsontang.com · 09/02/2025

What are you trying to do? uv kind of assumes you want to use venv for stuff and know how to activate those or use an IDE that can detect it. If you want some global python install then you have stuff like conda and pyenv

100

Nelson Tang @nelsontang.com · 06/02/2025

This whole time I’ve been learning Bayesian inference to detect biased coins and it turns out you can just look at it instead

000

Nelson Tang @nelsontang.com · 06/02/2025

You can run R in quarto and source() what you need I think

120

Nelson Tang @nelsontang.com · 05/02/2025

I guess I don't have a specific text I can share - you probably already have this, but the chapter in probml2 (Probabilistic ML - Advanced Topics) by Kevin Murphy that covers them as an intro to SSM is my other go-to probml.github.io/pml-book/boo...

probml.github.io

110

Nelson Tang @nelsontang.com · 05/02/2025

And I like this paper, discusses HMM and state space models and Bayesian networks - mlg.eng.cam.ac.uk/zoubin/paper...

mlg.eng.cam.ac.uk

110

Nelson Tang @nelsontang.com · 04/02/2025

It’s important to budget time for yak shaving

010

Nelson Tang @nelsontang.com · 04/02/2025

Now you too can refine macrodata! lumon-industries.com

lumon-industries.com

Lumon Industries

020

Nelson Tang @nelsontang.com · 04/02/2025

The quiet posters feed seems to be the best bet, otherwise everything just gets drowned out with people reacting to news

010

Reposted by Nelson Tang

Tyler Hillery @tylerhillery.com · 01/02/2025

Friendly reminder, don't forget to prune your uv cache

2123

Nelson Tang @nelsontang.com · 31/01/2025

Maybe an initial assignment could be around critiquing the LLM output based on something the student knows well or is an expert in. Like how you listen to certain podcast hosts and once you hear them talking about your field of expertise you realize they have no idea what they’re talking about

031

Nelson Tang @nelsontang.com · 31/01/2025

I wonder if, in addition to learning about a specific domain and their problems, you could teach students to critique a LLM’s output?

100

Nelson Tang @nelsontang.com · 31/01/2025

The analysts are motivated because they have immediate problems and these tools provide solutions and they learn skills at the same time. I don’t think it absolves them entirely of having to learn a little coding but LLM as coding mentor saves a ton of time for instructors

200

Nelson Tang @nelsontang.com · 31/01/2025

I don’t know what it would look like prior to entering the workforce, but arming working traditional excel-based analysts with the ability to solve their everyday problems with these new tools (in this case, writing Python with LLM help) works well.

100

Nelson Tang @nelsontang.com · 28/01/2025

time for a hierarchical model?

050

Nelson Tang @nelsontang.com · 17/01/2025

Wait there’s a payoff??

100

Nelson Tang @nelsontang.com · 17/01/2025

so glad they decided to make juliaup part of the official install process

120

Nelson Tang @nelsontang.com · 16/01/2025

Dev logs are new to me but they sound like a great idea. Here a Reddit user shares their format for keeping a dev log and also using it for some self care: www.reddit.com/r/Experience...

reddit.com

From the ExperiencedDevs community on Reddit: The way you treat yourself is important.

Explore this post and more from the ExperiencedDevs community

020

Nelson Tang @nelsontang.com · 09/01/2025

No commercials on YouTube alone makes it worth it! Music is just a great bonus

000

Nelson Tang @nelsontang.com · 09/01/2025

A few weeks back one of my primary stakeholders asked “why does it take so long to clean data? Shouldn’t it be instant?” And all the possible responses got jammed up in the doorway of my mind I think I blacked out

030

Nelson Tang @nelsontang.com · 09/01/2025

Too much work to block so I use a lot of mute words/tags and that seems to help? I haven’t seen much of them except when they follow me out of the blue but I assume they’ll fall off eventually if I don’t follow back

010

Nelson Tang @nelsontang.com · 27/12/2024

As a DS who works remote, this feels an elegant way of quickly spinning up a DAG and collaborating, brainstorming, or teaching without a whiteboard. It's pretty enough to drop into a stakeholder deck and it's got a generous free 'explorer' tier. Definitely belongs in the quiver

170

Reposted by Nelson Tang

travis gerke @travisgerke.bsky.social · 26/12/2024

In case any of you #rstats / #quarto folks are updating sites as a holiday task, TIL a slick way to get the Bluesky logo in: the iconify extension does it with: {{< iconify fa6-brands bluesky >}} github.com/mcanouil/qua... Also has academicons, etc 🎉

3757

Nelson Tang @nelsontang.com · 21/12/2024

On Gathering the Data: "This couldn’t be easier, as most of us are in many meetings on a daily basis. We recorded the arrival times of individuals for 10000 of our own meetings, which took us about three days."

120

Nelson Tang @nelsontang.com · 21/12/2024

A little late to the party here, but just read a fascinating (totally real) paper on applied gaussian processes. Titled "On the tardiness of coworkers and how to exploit it": jabde.com/2023/02/12/c...

jabde.com

On The Tardiness Of Coworkers And How To Exploit It - Journal of Astrological Big Data Ecology

In this groundbreaking research, we determine the optimum amount of time to show up late to a meeting by arbitrarily assigning numbers to things

2113

Nelson Tang @nelsontang.com · 14/12/2024

That article gave me anxiety

010

Nelson Tang @nelsontang.com · 14/12/2024

From a supply/demand pov I don’t think it’s super common to have a sales org that’s willing + able to pay for their own data folks and to find data folks who excel at working with non technical stakeholders.

110

Nelson Tang @nelsontang.com · 07/12/2024

Don’t think I’ve ever seen a better intro to state space models. Bonus points for interactivity!

140

Nelson Tang @nelsontang.com · 02/12/2024

Handy helper function for generating a dataset with causal relationships and visualize the DAG. This was one of my last bookmarks I had from the other place, glad to have found that @robdonnelly47.bsky.social reposted here.

060

Nelson Tang @nelsontang.com · 01/12/2024

+1 to this exact example - saw this work well on a data science team in which stakeholders got to bid based on business value and the data scientists put in bids based on feasibility

050

Nelson Tang @nelsontang.com · 30/11/2024

Jane Street is hosting a market data forecasting competition on Kaggle: www.kaggle.com/competitions...

kaggle.com

Jane Street Real-Time Market Data Forecasting

Predict financial market responders using real-world data.

041

Reposted by Nelson Tang

Paul Goldsmith-Pinkham @paulgp.com · 30/11/2024

Economists love using linear regression to estimate treatment effects — it turns out that there are perils to this method, but also amazing perks Come with me in this 🧵 if you want to learn about our now-published paper "Contamination Bias in Linear Regressions!" 1/ (Twitter rerun!)

18403125

Nelson Tang @nelsontang.com · 29/11/2024

In my house we call that a sad choice

010

Nelson Tang @nelsontang.com · 28/11/2024

Aside from making learning a part of your 9-5, I don’t have a great solution. You probably have to trade off something else (sleep, hobbies, partner time, child time) and accept that learning is going to take a slow, steady progress spread out over a long time

010