R demo | How to visualize models Part 2 | non linear, logistic, multinomial, mixed effects, survival

Why visualize model results? Because: "A picture is worth a thousand words"!

The goal is the same as in the previous video: learn minimum code for maximum results!

Here we'll see the visualisation of the non linear, logistic, multinomial, mixed effects, mixed effects non-linear and survival models, including Kaplan-Meier curves, exponential parametric models and even Cox Proportional Hazard models.

Moreover, we'll not only see how to do easy post-hocs again, but will learn how to visualize all the p-adjusted p-values of the pairwise comparisons (contrasts) on the same color-coded plot.

Visualizing model results is the best kind of data visualization, because such data is already processed and delivers even more valuable insights than the row data.

Idea on new video topics and any feedback are always welcomed!!!

Enjoy :)

Music by Vincent Rubinetti
Download the music on Bandcamp:
Stream the music on Spotify:

Enjoy! 🥳

[OnLyhereAlone]

I watched this series again and learned a lot more yet again. I wish i could give a second thumbs up. Thanks for the wonderful work.

[alexiscanari]

Thank you for sharing!! could you please specify if there are differences when we work with glmer for logistic mixed effects model? I was wondering particularly how to check this type of model. thank you!

[statbipin]

Hey Yury! How can we overlay exponential, weibull, etc. distributions over Kaplan-Meier Curve in R?

[chacmool]

Does the 'performance' package offer diagnostics for Logistic Regression? Can I pass a glm() object to its check_model() function?

[arturocdb]

Incredible usefull!, thank you very much!, great work!.

[shahfahadalishah]

Thanks sir very informative. Sir if possible then made video on structure equation modeling...

[Joao-gqwe]

Thank you for sharing your video! The explanation was very clear! I have a question: Is there a package to run multinomial model with random effects? I see that it isn't possible with nnet.

[mayurwabhitkar]

so i tried the check model function on this script
###Multiple non linear models with interactyion
m<-lm(log(salary)~poly(age, 2)*health, dd)
plot(allEffects(m))
instead of showing me the multicollinearity and non normality of residuals the output on my screen shows me a different prediction the log values has not been transformed
any idea what seems to be the problem

[yunes]

Really appreciate it 🙏 that what I was looking for 😻

[hooldhooldy]

Is it possible to visualise ordered probit model?

[siriyakcr]

How about Bayesian statistics and estimate in rstan, rstanarm