Discover and read the best of Twitter Threads about #julialang
Most recents (24)
Why are biologists adopting #julialang #sciml? Performance, metaprogramming, and the development of new abstractions are improving software tools for #computationalbiology #systemsbiology #bioinformatics. Check out this new paper in Nature Methods!
nature.com/articles/s4159…
nature.com/articles/s4159…
In this we detail how #julialang's core compute model gives faster code, with a detailed calculation of the effects of the #python interpreter and kernel launching costs on simulation performance. It's pretty cool how one can pen and paper calculate the 100x expected difference. 
Julia's ecosystem has a complete set of tools for mathematical modeling (#sysbio), #bioinformatics, #machinelearning, and #datascience which we contextualize in the field of biology.
🧹 Tidier.jl v0.7.1 is now on the #JuliaLang registry.
What’s new?
- drop_na()
- lag() and lead() - re-exported from ShiftedArrays.jl
- Bugfix to ntile() if all values are missing
Thanks to @KriseScheuch for feature suggestions!
github.com/kdpsingh/Tidie…
What’s new?
- drop_na()
- lag() and lead() - re-exported from ShiftedArrays.jl
- Bugfix to ntile() if all values are missing
Thanks to @KriseScheuch for feature suggestions!
github.com/kdpsingh/Tidie…
One interesting thing is that lag() and lead() take in a vector and return a vector (similar to ntile).
This means that these functions *should not* be auto-vectorized. So in addition to re-exporting, they are included on the package’s do-not-vectorize list.
This means that these functions *should not* be auto-vectorized. So in addition to re-exporting, they are included on the package’s do-not-vectorize list.
In the near future, we will provide a mechanism to edit the package’s do-not-vectorize list of functions.
For frequently used functions, this means you won’t have to use the tilde-prefix notation to call them.
See here for details on auto-vectorization: kdpsingh.github.io/Tidier.jl/dev/…
For frequently used functions, this means you won’t have to use the tilde-prefix notation to call them.
See here for details on auto-vectorization: kdpsingh.github.io/Tidier.jl/dev/…
#sciml #machinelearning in chemical engineering using prior scientific knowledge of chemical processes? New paper: we dive deep into using universal differential equation hybrid models and see how well gray boxes can recover the dynamics.
arxiv.org/abs/2303.13555 #julialang
arxiv.org/abs/2303.13555 #julialang
For learning these cases, we used neural networks mixed with known physical dynamics, and mixed it with orthogonal collocation on finite elements (OCFEM) to receive a stable simulation simulation and estimation process.
We looked into learning reaction functions embedded within diffusion-advection equations. This is where you have spatial data associated with a chemical reaction but generally know some properties of the spatial movement, but need to learn the (nonlinear) reaction dynamics
🧹Tidier.jl 0.6.0 is available on the #JuliaLang registry.
What’s new?
- New logo!
- distinct()
- n(), row_number() work *everywhere*
- `!` for negative selection
- pivoting functions are better
- bug fixes to mutate() and slice()
Docs: kdpsingh.github.io/Tidier.jl/dev/
A short tour.
What’s new?
- New logo!
- distinct()
- n(), row_number() work *everywhere*
- `!` for negative selection
- pivoting functions are better
- bug fixes to mutate() and slice()
Docs: kdpsingh.github.io/Tidier.jl/dev/
A short tour.
If you use distinct() without any arguments, it behaves just like the #rstats {tidyverse} distinct().
It checks if rows are unique, and returns all columns just as you would expect.
It checks if rows are unique, and returns all columns just as you would expect.
If you use distinct() with arguments as shown here, then it returns all columns for unique values of the supplied column.
This is slightly different behavior than {tidyverse} distinct(), but I kind of like it. Can easily pair this with select() to mimic dplyr behavior.
This is slightly different behavior than {tidyverse} distinct(), but I kind of like it. Can easily pair this with select() to mimic dplyr behavior.
"The Flaws of Inheritance" by @CodeAesthetic1 is beautiful, as always.
Problem though, is that none of the things discussed in the video have anything to do with inheritance
Time for a 🧵 on the most mind-bending construction in mainstream programming languages
Problem though, is that none of the things discussed in the video have anything to do with inheritance
Time for a 🧵 on the most mind-bending construction in mainstream programming languages
@CodeAesthetic1 Some context: I was talking to Norman Ramsey a few months ago about his new book. We started talking about objects, and he told me he barely covers them. Why? "Objects are not an undergrad topic"
Inheritance is probably the hardest part of the theory of objects.
If I say "X doesn't understand inheritance," read it in the same tone as "X doesn't understand quantum Chu algebras."
Here's how the late William Cook defined it
If I say "X doesn't understand inheritance," read it in the same tone as "X doesn't understand quantum Chu algebras."
Here's how the late William Cook defined it
My lab is moving to #JuliaLang, and I’ll be putting together some R => Julia tips for our lab and others who are interested.
Here are a few starter facts. Feel free to tag along!
Julia draws inspiration from a number of languages, but the influence of R on Julia is clear.
Here are a few starter facts. Feel free to tag along!
Julia draws inspiration from a number of languages, but the influence of R on Julia is clear.
Let's start with packages.
Like R, Julia comes with a package manager that can be used to install pkgs from within the console (or REPL). The Pkg package isn't automatically imported in Julia but it's easy to do.
Both are different from Python's command line approach to pkgs.
Like R, Julia comes with a package manager that can be used to install pkgs from within the console (or REPL). The Pkg package isn't automatically imported in Julia but it's easy to do.
Both are different from Python's command line approach to pkgs.
Julia natively takes pkg management much further than R. Want to install a package from GitHub? Easy, just add a url argument to the add function.
Pkg.add(url = "github.com/kdpsingh/TidyT…")
Pkg.add(url = "github.com/kdpsingh/TidyT…")
Differentiable programming (dP) is great: train neural networks to match anything w/ gradients! ODEs? Neural ODEs. Physics? Yes. Agent-Based models? Nope, not differentiable... or are they? Check out our new paper at NeurIPS on Stochastic dP!🧵
arxiv.org/abs/2210.08572
arxiv.org/abs/2210.08572
Problem: if you flip a coin with probability p of being heads, how do you generate a code that takes the derivative with respect to that p? Of course that's not well-defined: the coin gives a 0 or 1, so it cannot have "small changes". Is there a better definition?
Its mean (or in math words, "expectation") can be differentiable! So let's change the question: is there a form of automatic differentiation that generates a program which directly calculates the derivative with respect to the mean?
Met @abisen recently who pointed me to @duckdb and its #julialang interface. It was easily the easiest database I installed and ran. Installation was just `Pkg.add("DuckDB")` and I was running the examples all within a minute.
juliahub.com/ui/Packages/Du…
juliahub.com/ui/Packages/Du…
The next step is implementing the Tables.jl API so that it just plugs into the rest of the Julia ecosystem.
Do you want to learn #RStats, but aren't quite sure where to start?
Myself and some great colleagues from @_bios2 wrote down our best tips for learning R on your own, out today in @PLOSCompBiol 🔓
✨Ten Simple Rules for Teaching Yourself R✨
journals.plos.org/ploscompbiol/a…
🧵:
Myself and some great colleagues from @_bios2 wrote down our best tips for learning R on your own, out today in @PLOSCompBiol 🔓
✨Ten Simple Rules for Teaching Yourself R✨
journals.plos.org/ploscompbiol/a…
🧵:
RULE 1: Prepare for a steep learning curve 📈
#RStats is a language, and as such, won't be picked up in a day. When learning R, remember that progress will be slow and you may have some growing pains. As you get more practice, you'll become better at picking up new techniques.
#RStats is a language, and as such, won't be picked up in a day. When learning R, remember that progress will be slow and you may have some growing pains. As you get more practice, you'll become better at picking up new techniques.
RULE 2: Take the time to read a book 📚
No matter how general or specific an R skill you want to
learn, there is probably a great book on that skill!
One great place to look is the "Books" section on the
@rstudio website:
rstudio.com/resources/book…, but plenty more exist, too!
No matter how general or specific an R skill you want to
learn, there is probably a great book on that skill!
One great place to look is the "Books" section on the
@rstudio website:
rstudio.com/resources/book…, but plenty more exist, too!
Here's why #rstats users should be happy about Rstudio changing its name to Posit. 🧵
1⃣ It's easier to sell Posit products to your organization. Even though Rstudio Teams is a really good development ecosystem for Python work, it's very hard to express that because Rstudio's brand association is mostly R.
As a result it's going to be easier for R users to justify purchasing the products because it becomes a lot clearer that its not an investment in a particular technology.
📢 Physics + GPs + inverse problems using #ProbabilisticNumerics 📢
At #ICML2022 we show that probabilistic ODE solvers are not just fast, but also useful for solving inverse problems! Joint work with Filip Tronarp and @PhilippHennig5. More below 🧵
At #ICML2022 we show that probabilistic ODE solvers are not just fast, but also useful for solving inverse problems! Joint work with Filip Tronarp and @PhilippHennig5. More below 🧵
The gist is: When doing inference with traditional ODE solvers we ignore their numerical error. But by being "probabilistic about the numerics", we can fit _both the ODE and the data jointly_! Which e.g. allows us to better learn parameters of oscillatory systems:
Paper: proceedings.mlr.press/v162/tronarp22a
Experiments: github.com/nathanaelbosch…
Code in #julialang: github.com/nathanaelbosch…
And if you got curious about probabilistic ODE solvers, there's more this ICML:
Experiments: github.com/nathanaelbosch…
Code in #julialang: github.com/nathanaelbosch…
And if you got curious about probabilistic ODE solvers, there's more this ICML:
1/ The Built-in Package Manager
With the Julia REPL open, press the right square bracket ] to drop into the 𝙿𝚔𝚐 REPL where you can install packages and manage Julia environments.
There's no need to install a separate package manager. It comes with every Julia install.
With the Julia REPL open, press the right square bracket ] to drop into the 𝙿𝚔𝚐 REPL where you can install packages and manage Julia environments.
There's no need to install a separate package manager. It comes with every Julia install.
The support for Julia environments is one of my favorite parts. Inside the 𝙿𝚔𝚐 REPL, type:
𝚊𝚌𝚝𝚒𝚟𝚊𝚝𝚎 /𝚙𝚊𝚝𝚑/𝚝𝚘/𝚎𝚗𝚟𝚒𝚛𝚘𝚗𝚖𝚎𝚗𝚝
to activate the environment at the path, or create a new one if it doesn't exist.
𝚊𝚌𝚝𝚒𝚟𝚊𝚝𝚎 /𝚙𝚊𝚝𝚑/𝚝𝚘/𝚎𝚗𝚟𝚒𝚛𝚘𝚗𝚖𝚎𝚗𝚝
to activate the environment at the path, or create a new one if it doesn't exist.
My turn to learn the new #DiD literature.
I estimated tried as many relevant methods as I could find on two (simple) DGPs.
All #STATA code here: github.com/azev77/Compare…
#EconTwitter what did I miss? (1/n)
I estimated tried as many relevant methods as I could find on two (simple) DGPs.
All #STATA code here: github.com/azev77/Compare…
#EconTwitter what did I miss? (1/n)
Thanks to @AsjadNaqvi @borusyak @XJaravel @jannspiess @kylefbutts @pedrohcgs @CdeChaisemartin @arindube I missed a bunch (sorry)
1: asjadnaqvi.github.io/DiD/ is maybe the best resource to learn these methods
1: asjadnaqvi.github.io/DiD/ is maybe the best resource to learn these methods
2: the world would be much better off if these methods were implemented in a modern/fast language like #JuliaLang
github.com/JuliaDiffinDif…
by @JunyuanChenEcon
3: why doesn't anyone compare the staggered DiD methods to #SCM?
I think I'll do that...
github.com/JuliaDiffinDif…
by @JunyuanChenEcon
3: why doesn't anyone compare the staggered DiD methods to #SCM?
I think I'll do that...
Today's #JuliaLang doodle with Javis.jl is recreating a classic...
Look closely at the circles. Are the curving at all?
Here's the code:
(I'm sure there's a better way to do this, I just haven't completely figured out Javis yet!)
Here's the code:
(I'm sure there's a better way to do this, I just haven't completely figured out Javis yet!)
I think disks work better than circles for this:
#COVID19mx 2021-11-30 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,887,873 || nuevos: 3,307 (+0.09%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,887,873 || nuevos: 3,307 (+0.09%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 40,630 || variación diaria: +2,572 (+6.76%)
Acumulados: 40,630 || variación diaria: +2,572 (+6.76%)
#COVID19mx 2021-11-29 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,884,566 || nuevos: 724 (+0.02%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,884,566 || nuevos: 724 (+0.02%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 38,058 || variación diaria: -51 (-0.13%)
Acumulados: 38,058 || variación diaria: -51 (-0.13%)
#COVID19mx 2021-11-28 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,883,842 || nuevos: 1,050 (+0.03%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,883,842 || nuevos: 1,050 (+0.03%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 38,109 || variación diaria: +108 (+0.28%)
Acumulados: 38,109 || variación diaria: +108 (+0.28%)
#COVID19mx 2021-11-27 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,882,792 || nuevos: 2,956 (+0.08%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,882,792 || nuevos: 2,956 (+0.08%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 38,001 || variación diaria: -174 (-0.46%)
Acumulados: 38,001 || variación diaria: -174 (-0.46%)
Learning a new programming language can be difficult.
Our goal with juliaacademy.com is to make that a little easier. Here is a quick thread highlighting some of my favorite (free) #JuliaLang courses 🧵
Our goal with juliaacademy.com is to make that a little easier. Here is a quick thread highlighting some of my favorite (free) #JuliaLang courses 🧵
For those with programming experience, @JaneHerriman 's course on an Intro to Julia is extremely popular. Jane is a great teacher and this course helps you build a solid #JuliaLang foundation: juliaacademy.com/p/intro-to-jul…
For those with perhaps less programming experience / comfort, Dr Henri Laurie's "Julia for Nervous Beginners" course is an excellent option: juliaacademy.com/p/julia-progra…
#COVID19mx 2021-11-16 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,847,243 || nuevos: 735 (+0.02%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,847,243 || nuevos: 735 (+0.02%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 35,450 || variación diaria: -2,853 (-7.45%)
Acumulados: 35,450 || variación diaria: -2,853 (-7.45%)
#COVID19mx 2021-11-15 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,846,508 || nuevos: 775 (+0.02%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,846,508 || nuevos: 775 (+0.02%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 38,303 || variación diaria: -219 (-0.57%)
Acumulados: 38,303 || variación diaria: -219 (-0.57%)
#COVID19mx 2021-11-14 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,845,733 || nuevos: 942 (+0.02%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,845,733 || nuevos: 942 (+0.02%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 38,522 || variación diaria: -504 (-1.29%)
Acumulados: 38,522 || variación diaria: -504 (-1.29%)
#COVID19mx 2021-10-08 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,714,392 || nuevos: 7,158 (+0.19%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,714,392 || nuevos: 7,158 (+0.19%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 95,277 || variación diaria: -2,981 (-3.03%)
Acumulados: 95,277 || variación diaria: -2,981 (-3.03%)
#COVID19mx 2021-10-07 [hilo] #YoTengoOtrasGráficas
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,707,234 || nuevos: 7,613 (+0.21%)
Casos #CONFIRMADOS (por fecha de reporte)
Acumulados: 3,707,234 || nuevos: 7,613 (+0.21%)
Promedio móvil de 7 días de nuevos casos confirmados: por fecha de #reporte (curva roja) versus fecha de inicio de #síntomas (curva azul, se omiten últimas 2 semanas por rezago en información).
Casos #ACTIVOS (confirmados los últimos 14 días)
Acumulados: 98,258 || variación diaria: -4,195 (-4.09%)
Acumulados: 98,258 || variación diaria: -4,195 (-4.09%)
