Modin Project

Speed up your pandas workflows by changing a single line of code ⚡

[Modin Project @modin_project]

You can also find us at our Builder's Zone booth throughout the day to chat about Modin and all things Open Source! 🛠️🔍

[Modin Project @modin_project]

Join us this afternoon to learn how you can build an open-source AI stack with Modin, @streamlit, #Arctic to power your LLM App! 🧑‍💻💡

🗓️ Thursday, June 6 ⏰ 2:30 PM - 2:50 PM PDT
📍 Builders Hub Theater (Basecamp, South)

#SnowflakeSummit #OpenSource #AI #LLM

[Modin Project @modin_project]

Attending #SnowflakeSummit? ❄️ Visit our booth at Builder's Hub to explore Modin, speak with Modin OSS contributors, and check out cool demos! 👋 See you there!
#python #datascience #machinelearning

[Doris Lee @dorisjlee]

📣 This year's @SnowflakeDB #DataCloudSummit features #DevDay on June 6, open to all developers. Get ready for an exciting line of speakers from Andrew Ng (Founder @DeepLearningAI), Lukas Biewald (CEO/Co-founder @weights_biases ), and more!

I'll be giving a Dev Day luminary talk on how you can build an entirely open source #AI stack, including @streamlit, @modin_project, and Arctic!

[Software Dev Tools @IntelDevTools]

Learn how #oneAPI tools including the Intel Distribution of @modin_project, and @XGBoostProject optimizations from Intel are used to process data on freshwater supplies to determine whether or not the water is safe to drink: https://t.co/Mkfo9CToat

[Modin Project @modin_project]

Lots of great stuff in the most recent Modin release (0.29.0), including two functions to make it easier to work w/ @dask_dev dataframes: to_dask, and from_dask.

Check out the to_dask / from_dask code: https://t.co/NzSP85gJo2

And the release notes: https://t.co/lpcrC0GqKx

[Modin Project @modin_project]

Check out more details in the PR here: https://t.co/BhHh7YxEPU

[Modin Project @modin_project]

💥Performance Improvement for .merge in Modin!💥

The result? H2O benchmark (500 MB) total execution dropped by 70% (28.64s -> 8.78s)

What changed?
Before: Gather the right df into one partition
After: Repartition so we can broadcast only row partitions of the right df

[Modin Project @modin_project]

Dmitry Chigarev at @IntelDevTools wrote a post about Modin: https://t.co/1cEFVCXRNK

He takes a 4-minute pandas workload & says:
"Modin could be a solution here, offering a drop-in replacement for Pandas and efficient parallel implementations of its API."

With Modin? 55 seconds

[Ponder @ponderdata]

If you're backed into a dark alley, facing a pack of ravenous NaN dogs, you'll want a sidekick like pandas dropna.

Drop rows with n missing values using thresh=n.

Drop entirely empty rows with how='all'.

Learn more from @__mharrison__: https://t.co/cUxVKDc6ec

#pandas #Python

[Ponder @ponderdata]

If you want to automatically parallelize your pandas code, you should check out open-source Modin: https://t.co/uN0uRLwUWv

@modin_project uses the pandas API, so all you need to do is change "import pandas as pd" to "import modin.pandas as pd"

#pandas #Python #opensource

[Ponder @ponderdata]

.@IntelDevTools saw a workload speedup of 4X with @modin_project (the scalable drop-in replacement for pandas) compared to pandas!

Read more from Intel here: https://t.co/Evk4qmqls8

#python #pandas #datascience

[reneboy garcia @eboygarcia]

@andradavulpee can this be done faster using @modin_project ?

[Ponder @ponderdata]

The Python Dataframe Interchange Protocol is a quiet hero 🦸‍♀️.

It makes it easy for libraries to accept many dataframes (Modin, Polars, Ibis, cuDF, Dask).

Read our encomium: https://t.co/LWOnZBI5IL

Thanks to Marco Gorelli & @ralfgommers for your work!

#python #dataframes

[Ponder @ponderdata]

Want to learn 🧠 about @PyTorch DataLoaders 🏗️ ?

In this article, we:
- Present some background on DataLoaders
- Discuss how @modin_project can now do batch data loading through the new ModinDataLoader
- Share benchmarks

Check it out: https://t.co/8RdQlUsQ0Y

#pytorch #python

[Ponder @ponderdata]

Check out this new @_odsc blog post on scaling #pandas! https://t.co/MCkToL4tlt

Written by @ponderdata CEO @dorisjlee, it discuss the @modin_project + Ponder revolution where you can speed up / scale your pandas workflows with almost no changes to your scripts / notebooks.

[Ponder @ponderdata]

Exciting gathering in Switzerland this weekend: The #Python DataFrame Summit!

It's a chance for DF library maintainers (@jorisvdbossche @datapythonista Dea María Léon + more, incl. @maheshvash for @modin_project) to discuss roadmaps, etc.

Thanks to @pandas_dev for organizing!

[Modin Project @modin_project]

RT @ponderdata: Missed the pandas sessions at the last @pydata Global event? Catch up now!

@rdurrani22's + @ahpondering's sessions on pand…

[Ponder @ponderdata]

Modin's been downloaded 10 million times.

Along the way, it's been incorporated into major projects by Intel, AWS, and Ponder.

Learn more about Modin's growth story in this @intel post: https://t.co/UB7fqHLK7g

#pandas #python #datascience @IntelDevTools

[Modin Project @modin_project]

MODIN HIT 10 MILLION DOWNLOADS! 🔥
https://t.co/heLck276Tq

That's lots of "import modin.pandas as pd," lots of large dataframes processed, lots of time saved -- time NOT spent laboriously parallelizing code.

Congratulations, and thanks, to all!

#python #pandas #datascience