tomathon
banner
tomathon.bsky.social
tomathon
@tomathon.bsky.social
developer by day, caffeine junkie by night | 10+ yrs of experience still making the same rookie mistakes
impressive. very nice. would love to see some Rust IO supoort for Hive, HF, and BigQuery tho
January 14, 2026 at 9:28 PM
Happy Thanksgiving!
November 28, 2025 at 1:29 AM
Love the json_normalize() function! Very useful, especially when paired with pydantic for structure validation 😎

Also, the 100 days is almost over! I’m going to miss these posts.
June 29, 2025 at 3:38 PM
Nice one. Really useful when importing JSON object lists into data frame format.
June 20, 2025 at 1:24 PM
It’s wild how often this happens
June 16, 2025 at 1:24 PM
Enum is a really useful data type and IMO largely underutilized. I use it a lot with sorting and filtering named month data (ie June > April, etc).
June 10, 2025 at 9:39 PM
Yes! This is something I’ve noticed for other packages in the past. Strange.
June 9, 2025 at 2:12 PM
Yeah you can use the ‘how = “horizontal”’ arg to stack them horizontally.
June 1, 2025 at 12:17 AM
Nice! I always just use pl.concat(), but this seems like a much more elegant solution.
May 31, 2025 at 8:25 PM
Oh nice, I didn’t know about this one!
May 26, 2025 at 2:07 PM
This one is very useful when working with APIs, converting your data frame an object that can be passed to the body of a request.
May 19, 2025 at 2:36 PM
yyyy-mm-dd is my favorite date format so I really appreciate Polars making it easy to work with
May 14, 2025 at 4:38 PM
This is really neat. I came across Hyndman’s original R version of this book years ago and I remember thinking how cool it would be to have a Python version.
May 7, 2025 at 1:07 PM
This is really neat! I usually do this by first creating a map dict object and then using the .replace_strict() function to map variables. It helps standardize mapping for new data. But this is a really good alternative when working with one table at a time!
May 6, 2025 at 2:03 PM
Very useful!
April 30, 2025 at 2:09 PM
Very useful! I often use this along with a ‘with’ statement to preserve the global config.

with pl.Config(thousands_separator=“,”):
print(df)
April 21, 2025 at 3:16 PM
Bluesky needs a boormarks option for posts like this. Super useful stuff!
April 16, 2025 at 3:52 AM
Really useful! Didn’t know this existed and will be using it going forward.

These posts are terrific!
April 14, 2025 at 2:14 PM
RF on its own is criminally underrated. But yes, I agree on XGBRF reg too.
April 8, 2025 at 2:31 AM
The .add() method is pure zen of python. I prefer it personally.
April 1, 2025 at 5:38 PM
These are really useful!
March 25, 2025 at 1:48 PM