Dr Charles T. Gray 🎵 ⚔️ (@cantabile) header
Dr Charles T. Gray 🎵 ⚔️ (@cantabile) profile picture

beat jedi 🎵⚔ posting datasci thoughts to linkedin

Frederiksberg, Danmark
Joined January 2008
Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)Photo by Dr Charles T. Gray 🎵 ⚔️ (@cantabile)
Scraping open video game players' data has been such a ride. So much fun presenting with friends from Unity at PyData Copenhagen last night. talk: youtube.com/live/6HhPumAORIY… Documentation for the project in a pkgdown site. softloud.github.io/supermetr…
1
5
688
Dr Charles T. Gray 🎵 ⚔️ retweeted
From music to mathematics and data visualisation: Taking a front seat on the wild ride that is the data revolution 🎢👾 womanthology.co.uk/from-musi… @cantabile #WomenInMathematics #WomenInMaths #WomenInData
2
3
285
Dr Charles T. Gray 🎵 ⚔️ retweeted
Around Copenhagen and curious about data analysis and/or #RStats? The CopenhagenR meetup is closing the spring season, on the 12th of June, with a ✨ super-interesting✨ presentation by @cantabile : Animating a melody as a mathematical object. 🎶 😊 meetup.com/copenhagenr-user-…
1
1
8
692
Dr Charles T. Gray 🎵 ⚔️ retweeted
I am a programmer, and am vocally critical of LLMs like ChatGPT; they're completely untrustworthy, rely heavily on unattributed data, and creating them is environmentally devastating. I regularly ask ChatGPT for coding advice anyway.
24
10
2
877
209,437
Dr Charles T. Gray 🎵 ⚔️ retweeted
Statistics, data science and machine learning walk into a bar. Bartender: table for one, I assume?
9
30
3
203
26,535
I've encountered some really varied expectations of a career in data science. Academics make the distinction between academia and industry, but, in my experience, there really hasn't been any difference in terms of work needed. Robust pipes, some models, mostly EDA in vis.
1
3
412
Disappointment in modelling folk who don't like the software eng side of datasci positions, I find this odd, I've always found data science to be intersectional between math and code, in terms of employment. I do find I need to be upfront about my lack of skill in backend.
120
Spending the day migrating our game events data. So satisfying to patch in a different data set at staging, and watch all the pipes run 😍
3
228
Dr Charles T. Gray 🎵 ⚔️ retweeted
The sweet spot of the Simpsons, maybe series 2 to series 9, is the greatest achievement in the history of mankind.
31
16
3
503
93,792
I have heated floors throughout my apartment; I have never known such luxury. Never been colder outside, just had a week of -6, but never been so warm at home. It's the little things 🥰
165
Dr Charles T. Gray 🎵 ⚔️ retweeted
We're looking for #dyspraxic adolescents in Devon to help us validate a newly-developed #VR screening tool. Participants get a MABC-2 screening, compensation for their time (£30), and the chance to play some cool newly-developed VR games. Sign up here! shorturl.at/amnSX
4
4
1,161
😯 there's a fivetran connector for Google drive. Amazing, we have sheets we share with stakeholders for labelling and now we can stage them at the utmost up of the upstream of our DBT pipes 😍
160
Thrill of working in Scandinavia doesn't go away 🥰❄️
5
376
"Seeing math in code is like hearing the song" 😍
I understand this frustration, but I have a slightly different take. Some people can read music and hear it as they read. Other people need to hear it played first before they can play it themselves. Seeing math in code is like hearing the song.
1
1
231
There are still some moments of poetry here, I'm finding, after being off social media for a few years - living in Europe is most diverting.
1
94
So, what would a meaningful effect be? Stats school: your stakeholders will advise Stakeholders: what's realistic under different conditions? We will adjust our expectations accordingly. 🤔 Exploratory data analysis is such a necessary first step so much of the time.
We often need larger sample sizes for accurate estimates than to claim there is an effect with low error rates. It is not unreasonable in research lines to first establish there is an effect large enough to matter, and only if this is so, estimate it with high accuracy.
1
2
2
3,291
Math is such a ride. Today I shipped an analysis of a new design feature to the producer of the game (new feature works). Never dreamt I'd get to do things like this.
10
506
My nisse send Danish good cheer and happy holidays
2
133
Oh how I love my street ❤️
5
243
For those of you who are still here, check out this open-source command line tool for reproducible cloud-based data engineering pipelines. 2 weeks in, and I'm evangelical. Many "They should have sent a poet" moments as I find new functionalities. github.com/dbt-labs/dbt-core
1
6
425
Oh, how I would have waxed lyrical in my dissertation, @rlmcelreath 🤣
1
1
171