Twitter | Search | |
Mike Kearney📊
Assistant Professor at and MU Informatics Institute. Interested in new media, partisanship, data journalism, data[["big"]] science, and .
3,356
Tweets
1,033
Following
3,374
Followers
Tweets
Mike Kearney📊 4h
Anyone interested in learning or teaching natural language processing (NLP), this is a fantastic resource: Speech and Language Processing (3rd ed. draft) by and
Reply Retweet Like
Mike Kearney📊 4h
Replying to @petemohanty
I agree...and it's not so sparse (like the un-logged data) you can actually see the distribution!
Reply Retweet Like
Mike Kearney📊 retweeted
Emily 🦔 Vietti 4h
Replying to @kearneymw
If logging you is wrong, I don’t wanna be right. ... I’m really sorry, I couldn’t help it.
Reply Retweet Like
Mike Kearney📊 6h
It feels wrong but I'm a fan of these log transformed x scales (achieved with ggplot2 via `scale_x_log10()` )
Reply Retweet Like
Mike Kearney📊 7h
The iPhone has been stubbornly protesting linear messaging for several months now. And the MSM hasn't said a word about it!
Reply Retweet Like
Mike Kearney📊 retweeted
Andrew Heiss 16h
wait we’re not really doing this whole arm the teachers thing are we?
Reply Retweet Like
Mike Kearney📊 retweeted
Jenny Bryan Feb 21
How to tackle your encoding problems in . Kevin has distilled what I am sure is many hours of human pain into this fine post for us 🙌
Reply Retweet Like
Mike Kearney📊 retweeted
Guy McHendry Feb 21
The thing about snowflakes is when enough of us get together we can shut things down. ❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️❄️
Reply Retweet Like
Mike Kearney📊 retweeted
Garrick Aden-Buie Feb 17
Some weekend fun. I was inspired by brilliant use of the USDA watercolors in his rstudioconf talk, so I created a pomological ggplot2 theme: ggpomological! 🍑🍇🍎📊
Reply Retweet Like
Mike Kearney📊 retweeted
Maëlle Salmon 🐟 Feb 21
My latest blog post for "Markdown based web analytics? Rectangle your blog"
Reply Retweet Like
Mike Kearney📊 retweeted
Matt Grossmann Feb 20
Clinton waged more ad hominem attacks than Trump in her ads; Trump & supportive groups were more likely to run contrast ads on policy; Clinton did not use a full range of message strategies or bolster her own image
Reply Retweet Like
Mike Kearney📊 Feb 20
I'm an extra dose of CO2 without the climate changing side effects!
Reply Retweet Like
Mike Kearney📊 retweeted
Pew Research Center Feb 20
We know that survey mode effects and data weighting aren’t on everyone’s short list of water-cooler conversation topics. That's why we made this video:
Reply Retweet Like
Mike Kearney📊 Feb 20
Replying to @sarhutch @IRE_NICAR
👏👏👏
Reply Retweet Like
Mike Kearney📊 Feb 20
Replying to @sarhutch @IRE_NICAR
It's a weird thing to plan but it seems about right...
Reply Retweet Like
Mike Kearney📊 Feb 20
Replying to @charlesminshew
lol
Reply Retweet Like
Mike Kearney📊 retweeted
G. Elliott Morris📈🤷‍♂️ Feb 20
Replying to @gelliottmorris
Democratic candidate in KY special election FLIPS SEAT by a 38% margin, in a district Trump won by 50%. This massive 86% swing is the biggest one so far and pushes average D swing since 2016 up to 13%. This is what a wave looks like
Reply Retweet Like
Mike Kearney📊 retweeted
Rebekah Tromble Feb 20
Replying to @seanjtaylor
A good thread, but this one gives me qualms. Digging through data is crucial, esp to combat ML's tendency to latch onto spurious features. But simply digging around in search of s'thing "interesting" can lead to the exact same problem. It's why we need theory-guided questions.
Reply Retweet Like
Mike Kearney📊 Feb 20
We basically posted the same tweet at the exact same time!
Reply Retweet Like
Mike Kearney📊 Feb 20
It's a good thread...but if we're going to remind the AI/ML enthusiasts of the value of empiricism, perhaps we should also remind them of the value of "theory" as well? Interacting with data is good. Interacting with and *understanding* data is better. CC
Reply Retweet Like