Julia! You're a life-saver! I'm currently working in the Ecological and Environmental Sciences industry with plans of getting into a Masters of Econometrics/Mathematics soon. This video (along with others) are great resources, and I can't thank you enough for putting them up! Just purchased your book on Amazon, hoping to devour it when I get my hands on it.
Julia, thank you so very much for these videos. They are incredibly helpful for someone new to R like myself who learns best by video demonstration and replication in RStudio.
I don't think I've done a screencast, but here is an example with a keras neural network: www.tidymodels.org/learn/models/parsnip-nnet/ And my recent book with Emil Hvitfeldt has a section on using keras with tidymodels packages: smltar.com/dloverview.html
It's been a while since I made this video, but I believe because I was preparing to train a model to predict whether a school was above/below the threshold, using that same factor variable.
Hi Julia, thank you for the video. I wonder why at 10:48, the coeffocient for Arkansas is positive and > 1? If 1 = above threshold, then this coefficient does not make sense... Or am I misunderstanding anything?
The outcome is the factor `mmr_threshold`, which has levels "Above" and "Below". Because R defaults to alphabetical order for factors, the model is predicting the probability of being below compared to being above. When I use the `predict()` function later, notice that I pick out the probabilities of being above.
@@JuliaSilge Thank you for your time of explanation! Wow I did not know such details as R defaults to aphabetical order for factor in function. I understand now. Thank you again & have a nice day!
Julia, This looks like a very interesting video. My problem, however, is that I am having a hard reading what is on the screen. It is small and also looks fuzzy. Perhaps you can make the font bigger or expand the screen. Regarding the fuzzyness maybe the light background is better than dark.
You might try a Chromium browser (Chrome, Brave) if you are having trouble with resolution. UA-cam videos tend to render better in those. Also, under the settings gear, make sure you have chosen HD 1080p quality.
Julia, I use Chrome and I am view these videos in youtube for clarity so that is not the problem. I also know that these settings are clearly not an issue when I watch other videos like Mike Marin (R and Stats) and Greg Martin (Global Health - he does some nifty R videos also). Same goes for the LaTeX lectures that Michelle Krummel has up and Derek Banas (he uses a dark background like you and his font is also a little bit fuzzy but size makes up for it). I am telling you this not because I am an isolated case - more I am likely not. I, for one, click on interesting R video topics and if I can't read the screen then I am gone. By making your text and or screen bigger I am sure you get more hits and the followers you deserve. Just a suggestion.
@@haraldurkarlsson1147 No worries -- I definitely appreciate feedback, especially about accessibility issues! In some of my early videos like this one, I know I did not have the size bumped up enough but have since changed how I record. 👍
I have a quick question that I hope you don't mind answering. I have been working with the rebus package (with stringr) in RStudio and wondered if there was a shortcut for %R% sequence similar to %>& (for a Mac). I have not seen one. I thought you might know as an RStudio employee.
Hi Julia, your series is awesome. But I was not able to install "rstanarm" package. It asks me "Do you want to install from sources the package which needs complication". I entered "Yes". When I load "rstanarm", it says "Error: package or namespace load failed" . What should I do?
rstanarm can be a bit difficult to install. Try again and say "no", which means you will install a binary instead of the source, which needs to be compiled.
@@JuliaSilge Thanks, Julia. I did what you said(installed the binary pack), but when I conduct "library(rstanarm)", this message appeared. Error: package or namespace load failed for ‘rstanarm’ in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so': dlopen(/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so, 6): Symbol not found: _EXTPTR_PTR Referenced from: /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so Expected in: /Library/Frameworks/R.framework/Resources/lib/libR.dylib in /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so In addition: Warning message: package ‘rstanarm’ was built under R version 4.0.2
@@JuliaSilge I installed a binary of rstan, but when I load this, this error message comes out. Error: package or namespace load failed for ‘rstan’ in dyn.load(file, DLLpath = DLLpath, ...): unable to load shared object '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so': dlopen(/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so, 6): Symbol not found: _EXTPTR_PTR Referenced from: /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so Expected in: /Library/Frameworks/R.framework/Resources/lib/libR.dylib in /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so
@@sophiej4605 Hmmmmm, sounds like you may need a bit more detailed help than just going back and forth in comments here. I suggest you look in detail at the instructions about installing rstan here: github.com/stan-dev/rstan/wiki/RStan-Getting-Started And then look at the Stan forums to folks with similar problems to see how they solved it: discourse.mc-stan.org/
You can install it from GitHub using devtools::install_github("juliasilge/silgelib"): github.com/juliasilge/silgelib However, you can use something like theme_minimal that comes with ggplot2 for a pretty similar look, but with default fonts. You will need to install the fonts on your computer as well. If you want some theming that is more out-of-the-box, check out the hrbrthemes package: cinc.rud.is/web/packages/hrbrthemes/
These screencasts are what the UA-cam bell icon was meant for. These, and David Robinson's :)
Julia! You're a life-saver! I'm currently working in the Ecological and Environmental Sciences industry with plans of getting into a Masters of Econometrics/Mathematics soon. This video (along with others) are great resources, and I can't thank you enough for putting them up!
Just purchased your book on Amazon, hoping to devour it when I get my hands on it.
These videos are the best part of my week. Thanks, Julia!
Julia, thank you so very much for these videos. They are incredibly helpful for someone new to R like myself who learns best by video demonstration and replication in RStudio.
This is awesome! I didn’t know about skimr- I like it.
Thanks for this video, Julia.
I love your videos so much, you're seriously the best. Do you ever use neural networks in tidymodels?
I don't think I've done a screencast, but here is an example with a keras neural network:
www.tidymodels.org/learn/models/parsnip-nnet/
And my recent book with Emil Hvitfeldt has a section on using keras with tidymodels packages:
smltar.com/dloverview.html
Thanks for putting together these tidymodels videos. I think tidymodels will be a big deal and your videos are ahead of the curve.
These videos are the absolute best.
You are amazing. I am learning a lot from your screencast. Thank you very much.
I love you Julia!!!!, you are the best!
Thank you Julia for this video. Quick one why did you choose to summarise using the factor variable mmr_threshold instead of mmr >95?
It's been a while since I made this video, but I believe because I was preparing to train a model to predict whether a school was above/below the threshold, using that same factor variable.
Hi Julia, thank you for the video. I wonder why at 10:48, the coeffocient for Arkansas is positive and > 1? If 1 = above threshold, then this coefficient does not make sense... Or am I misunderstanding anything?
The outcome is the factor `mmr_threshold`, which has levels "Above" and "Below". Because R defaults to alphabetical order for factors, the model is predicting the probability of being below compared to being above. When I use the `predict()` function later, notice that I pick out the probabilities of being above.
@@JuliaSilge Thank you for your time of explanation! Wow I did not know such details as R defaults to aphabetical order for factor in function. I understand now. Thank you again & have a nice day!
Thanks for another screencast! :)
Julia, This looks like a very interesting video. My problem, however, is that I am having a hard reading what is on the screen. It is small and also looks fuzzy. Perhaps you can make the font bigger or expand the screen. Regarding the fuzzyness maybe the light background is better than dark.
You might try a Chromium browser (Chrome, Brave) if you are having trouble with resolution. UA-cam videos tend to render better in those. Also, under the settings gear, make sure you have chosen HD 1080p quality.
Julia, I use Chrome and I am view these videos in youtube for clarity so that is not the problem. I also know that these settings are clearly not an issue when I watch other videos like Mike Marin (R and Stats) and Greg Martin (Global Health - he does some nifty R videos also). Same goes for the LaTeX lectures that Michelle Krummel has up and Derek Banas (he uses a dark background like you and his font is also a little bit fuzzy but size makes up for it). I am telling you this not because I am an isolated case - more I am likely not. I, for one, click on interesting R video topics and if I can't read the screen then I am gone. By making your text and or screen bigger I am sure you get more hits and the followers you deserve. Just a suggestion.
@@haraldurkarlsson1147 No worries -- I definitely appreciate feedback, especially about accessibility issues! In some of my early videos like this one, I know I did not have the size bumped up enough but have since changed how I record. 👍
Thanks for doing these videos - the content is first class. Keep up the good work.
I have a quick question that I hope you don't mind answering. I have been working with the rebus package (with stringr) in RStudio and wondered if there was a shortcut for %R% sequence similar to %>& (for a Mac). I have not seen one. I thought you might know as an RStudio employee.
Great video. Thanks 👍
Hi Julia, your series is awesome. But I was not able to install "rstanarm" package. It asks me "Do you want to install from sources the package which needs complication". I entered "Yes". When I load "rstanarm", it says "Error: package or namespace load failed" . What should I do?
rstanarm can be a bit difficult to install. Try again and say "no", which means you will install a binary instead of the source, which needs to be compiled.
@@JuliaSilge Thanks, Julia. I did what you said(installed the binary pack), but when I conduct "library(rstanarm)", this message appeared.
Error: package or namespace load failed for ‘rstanarm’ in dyn.load(file, DLLpath = DLLpath, ...):
unable to load shared object '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so':
dlopen(/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so, 6): Symbol not found: _EXTPTR_PTR
Referenced from: /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so
Expected in: /Library/Frameworks/R.framework/Resources/lib/libR.dylib
in /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so
In addition: Warning message:
package ‘rstanarm’ was built under R version 4.0.2
@@sophiej4605 OK, trying installing a binary of rstan, i.e. `install.packages("rstan")` and then installing the binary for rstanarm.
@@JuliaSilge I installed a binary of rstan, but when I load this, this error message comes out.
Error: package or namespace load failed for ‘rstan’ in dyn.load(file, DLLpath = DLLpath, ...):
unable to load shared object '/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so':
dlopen(/Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so, 6): Symbol not found: _EXTPTR_PTR
Referenced from: /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so
Expected in: /Library/Frameworks/R.framework/Resources/lib/libR.dylib
in /Library/Frameworks/R.framework/Versions/4.0/Resources/library/rstan/libs/rstan.so
@@sophiej4605 Hmmmmm, sounds like you may need a bit more detailed help than just going back and forth in comments here. I suggest you look in detail at the instructions about installing rstan here:
github.com/stan-dev/rstan/wiki/RStan-Getting-Started
And then look at the Stan forums to folks with similar problems to see how they solved it:
discourse.mc-stan.org/
What about the logistic regression assumptions?
Excellent!
Woah ! thank you so much Julia !
Hello! How can I install the silgelib library? Thanks in advance. Wonderful videos. Regards
You can install it from GitHub using devtools::install_github("juliasilge/silgelib"): github.com/juliasilge/silgelib
However, you can use something like theme_minimal that comes with ggplot2 for a pretty similar look, but with default fonts. You will need to install the fonts on your computer as well.
If you want some theming that is more out-of-the-box, check out the hrbrthemes package: cinc.rud.is/web/packages/hrbrthemes/
@@JuliaSilge Thank you very much! you are very kind, thank you very much for sharing your knowledge!
Awesome videos julia
I can remember taking a (oral) polio vaccine in grade school(1-6 grade).
I was in primary school (UK version of grade school) and I remember getting an injection lol ... Injections have traumatised me since that day. 😂🥲
Thank you so much!!!
Thanks for the video, do you share your source code anywhere?
I almost always do, but this one didn't make it into blog form. You can find my tidymodels blog posts here: juliasilge.com/category/tidymodels/
😀😀😀,thanks
Tks for share!
Neat.
Arkansas is last in everything :D