August 6, 1992 looks like. Oakland vs Texas

Guess it's time for a new laptop then. Darn 😃

That sounds way more efficient than what I did. I actually struggled a lot with the size - working on a 10-year-old 16gb RAM thinkpad but with only like 3gb free on the hard drive.

Duckdb kept spilling to disk and then I'd get "out of memory" errors. At least I think that's what was happening

These are amazing. I'm guessing I missed then because I specified only "verb" part of speech when lemmatizing?

This was my first time using Github Copilot. I had a pretty good idea of the approach I wanted to take so I was mainly leaning on it for fleshing out the syntax. I found myself correcting quite a bit of the SQL it wrote, but overall I was very impressed!

The results:
• Tom Browning (present participle) vs. John, Michael, and Jarvis Brown (36 PA in the 1980s)
• Barret Browning vs. Dominic Brown (2 PA in 2012)
• Kelly Downs vs. Brian Downing (3 PA in 1992)
• Jace Fry vs. Max Fried (1 PA in 2019)

And that's it! Open to any challenges or corrections.

My process and tools:
• Retrosheet's full parsed play-by-play dataset (~15M records, ~200K games)
• DuckDb to load .csv and generate unique batter/pitcher name combos
• NLTK to apply verb lemmatization, keeping only the pairs with matching lemma
• Claude 3.5 + GH Copilot + vibes

My favorite that I’ve seen is the one @beeonaposy.bsky.social built and shared.

www.caitlinhudon.com/posts/2020/0... An Intake Form for Data Requests — Haystacks by Caitlin Hudon

@itsnotaboutthecell.com looks like it goes way back!

dax.tips/2019/05/14/a...

Shame on me for using a janky workaround…but at the same time it’s worked for ~5 years this way with no issues. I feel like I might’ve learned it from BIElite back in the day.

And no warning that I saw (although I very well could have missed something)

How my Monday started:

“ Measures” (leading space) is evidently no longer a supported table name, as of 2.140.1078.0, preventing us from opening models in Desktop.

The release cadence seems to have slowed considerably. Unsure if that’s related

I agree, but also “not everything that counts can be counted, and not everything that can be counted counts”.

Aren’t there some things that we should do just because they make sense, even if we have no idea how or whether we’ll be able to directly measure their impact?

Thanks for highlighting this book. Binge-listened (read?) it over the long weekend, inspired by this post. Really eye-opening perspective.

A lot of Roman era columns were as well, I think.

Took this photo in Pompeii last month.

Not natively, but there is a “workaround”. Although evidently it’s been having some issues as of late.

bsky.app/profile/josh...

The common "solution" is just to save a copy of the dashboard as a pdf each month lol. Full circle

There are several of us here!