Sabitlenmiş Tweet
Haven't been super active on twitter lately anyways but figured I'd give Mastodon a try -- I'm at @AliciaSchep@vis.social
English
Alicia Schep
159 posts
Alicia Schep
@AliciaSchep
data scientist with broad interests, including data visualization, healthcare, trees, dogs, etc. pronouns: she/her. Maybe at @[email protected]
Katılım Temmuz 2013
501 Takip Edilen553 Takipçiler
Alicia Schep retweetledi
1/2 Delighted to share a python package pychromVAR that implemented the idea from the chromVAR paper published in by @AliciaSchep et al for inferring TF variability from scATAC-seq data.
English
@AliciaSchep Really cool. Can you send a pull request to add vegabrite to vega.github.io/vega-lite/ecos…?
English
Following the update of R package `vegawidget` to Vega-Lite 5, have updated the package formerly known as `vlbuildr` to match and renamed it to `vegabrite`! Built out an example gallery show-casing variety of plots that can be made at vegawidget.github.io/vegabrite/arti…
English
Alicia Schep retweetledi
📣 #rstats
very happy to announce CRAN release of 📦 vegawidget 0.4.1
👉 renders charts using Vega-Lite 5 *or* 4
👉 uses {V8} package to create images
vegawidget.github.io/vegawidget/
English
Nice post but I think it showcases how SQL makes some tasks disproportionately gnarly. Recently had to compute similar metrics in dbt and opted for a javascript UDF, which felt more modular. But in e.g. ruby you'd just use the business_time gem and save a lot of headaches!
Jason Ganz@jasnonaz
New up on the dbt Developer blog - @_daveconnors3 walks through an important business metric to us at dbt Labs, Time on Task or how to measure business hours in SQL - both how to calculate it and critically what the business value of measuring this is. docs.getdbt.com/blog/measuring…
English
@drbridgewater @teej_m I would agree with you but also think humility is part of being a good applied statistician 😀 Including humility around limits of statistical approaches and importance of domain knowledge!
English
@teej_m Nah, computational biology and genetics .../cc @AliciaSchep
English
Comparing physics to linear regression (and picking regression as the winner) is the weirdest take I've seen here in a while.
Also linear regression got a lot more useful with computers (based on physics).
Dr Kareem Carr@kareem_carr
Controversial opinion: Physics is overrated. Statistics gave us linear regression which allows us to make simple extrapolations from almost any kind of numerical data. This ability has contributed massively to human progress. Much more useful that some silly physics equation.
English
Grateful to be represented in our state government by Assembly Member Wicks. Her thread highlights so powerfully why the ongoing onslaught on abortion rights is horrifying
Buffy Wicks@BuffyWicks
How do I begin to unpack how deeply disturbed I was by yesterday’s Supreme Court arguments? Justice Sotomayor spoke for so many of us with this question. She spoke for me. In September, I had a miscarriage requiring an emergency abortion procedure — (tw // miscarriage)
English
Agree that a lot of BI tools are great for rapid exploration! But that's also why I think they can be a trap: making dashboards might be easy, but maintaining them and ensuring the integrity of results over time is a whole different story.
Hamel Husain@HamelHusain
Super uncool opinion: For adhoc exploration of data, I think something like Tableau is far superior vs. writing a bunch of maltplotlib code.
English
@crcgrubbsd Sorry don't have bandwidth right now! Sounds like an interesting topic though
English
@AliciaSchep I have a book proposal on Healthcare Data Wrangling in R and wondered if you would be interested in doing a review. I can pay you in free books. Interested?
English
Really cool opportunity to have an impact on the ethics & reproducibility of research by developing alternatives to fetal bovine serum👇
English
Thank you to Joe Biden and Kamala Harris and every American who voted for a better future. So much work still to do, but today I'm feeling hopeful and grateful 🎉🇺🇲❤️
English
Alicia Schep retweetledi
The economy.
Climate change.
Health care.
Civil rights.
Racial justice.
The U.S. Supreme Court.
Our democracy.
They’re all on the line. Vote. iwillvote.com
English
@domoritz @FirasM @dataandme @W_R_Chase @hadleywickham @thomasp85 @vega_vis @ijlyttle I would describe vegawidget more as a building block than an mvp; it abstracts going from a json spec to a rendered visualization in R so other packages can focus on going from R code to a spec (b/c can then use vegawidget for rendering).
English
@FirasM @dataandme @W_R_Chase @hadleywickham @thomasp85 @vega_vis @ijlyttle @AliciaSchep I think if you want a seamless @vega_vis integration with R, you have to write the bindings directly for R. Vegawidget is an awesome MVP to learn from.
English
I interviewed @hadleywickham about the history of R and it's role in data visualization! Big thanks to Hadley for giving this interview! I thoroughly enjoyed it and think there's some pretty great insights here. #rstats medium.com/nightingale/da…
English
Alicia Schep retweetledi
obviously there is no accountability in this broken, ruined society but in a modestly functional world no one responsible for inflicting this man on the country would be able to show their faces in public again
English
@Jemus42 The internal package is only needed to build the main package and not to run it, so I think it would get left out of what is sent to cran... In other words cran shouldn't have to know about it!
English
Alicia Schep retweetledi
This is a question of values. Do we want to invest in Mayor Bloomberg, or in an entire generation of young people? #DemDebate
English
Alicia Schep retweetledi
A year ago in Nature Biotechnology, Becht et al. argued that UMAP preserved global structure better than t-SNE. Now @GCLinderman and me wrote a comment saying that their results were entirely due to the different initialization choices: biorxiv.org/content/10.110…. Thread. (1/n)
English
Alicia Schep retweetledi
New post, about code generation in R packages, inspired by a thread started by @MilesMcBain!
blog.r-hub.io/2020/02/10/cod…
"if you write three embarrassingly similar functions… write code to generate their code for you?" 😎
#rstats
English