My last lecture for #psy6136, Categorical Data Analysis,
"The Last Waltz" summarizes much of the content &points to
areas I could cover in a longer course (Bayesian, multilevel data, LCA, ...)
[Hardly anyone got the Last Waltz -- The Band reference 🥲]
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
Lecture 9 for my course in Categorical Data Analysis-- Count Data Models
📋 Materials: friendly.github.io/psy6136/#GLM...
🎞️ Slides: friendly.github.io/psy6136/lect...
Less likely to need flashing like if they come with effective plots. Interpreting coefficients? OK, but so old school
#psy6136 lecture this week on Count Data models.
friendly.github.io/psy6136/
Overiew of logistic regression models for polytomous responses: Proportional odds model, nested logit model, multinomial logit model
Plot of predicted probabilities of levels of womens' labor force participation for a nested logit model
#rstats #dataviz #psy6136
Lecture 6 for my course in Categorical Data Analysis-- Extending Logistic Regression (polytomous responses)
Materials: friendly.github.io/psy6136/#Log...
Slides: friendly.github.io/psy6136/lect...
Title slide for lecture on Deep Questions of Data Visualization. It shows three types of stimuli (conch pie charts, granola bar charts and shrimp start charts) in an underwater scene with icons for sharks, turtles and stringrays
Slide illustrating the design of an experiment comparing conch pie charts, granola bar charts presented to sharks as subjects
Extended experiment, adding shrimp start charts as stimuli and stingrays as subjects
The author, celebrating the Red Stripe Award from the Deep Questions Research Institute
#rstats #dataviz #psy6136
BONUS lecture for my course in Categorical Data Analysis-- Deep Questions of Data Visualization
First ever study of graphical preferences among marine animals!! 🦈🐢🐟
WINNER of the Red Stripe Award, 2026!! 🎖️🎖️
Slides: friendly.github.io/psy6136/lect...
A slide on the age-old question. One panel shows two buttons labelled Plot first? Fit first? The second panel shows a person wiping his head, with the caption, "I can plot-fit-plot"
I'm revising a lecture on extensions of logistic regression for my #psy6136 #CDA course and thought I'd share this slide on t he age-old question:
PLOT FIRST?
FIT FIRST?
It captures a course theme that data analysis is an iterative process, involving a loop of graphing and model fitting.
#rstats #dataviz #psy6136
Lecture 6 for my course in Categorical Data Analysis-- Logistic Regresson
Materials: friendly.github.io/psy6136/#Log...
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
Lecture 5 for my course in Categorical Data Analysis-- Correspondence Analysis
Materials: friendly.github.io/psy6136/#Cor...
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
Lecture 4 for my course in Categorical Data Analysis-- Loglinear models and Mosaic displays -- one of my fav topics
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
I added a new page to my CDA course web site, friendly.github.io/psy6136/R-fu... giving an overview of the R functions in various packages related to the course topics, a kind of extended cheatsheet.
#rstats #dataviz #psy6136
vcdExtra::color_table() is coming along nicely. Now, display either cell frequencies or residuals, with shading representing associations that remain after fitting some model.
#rstats #dataviz #psy6136
Here's an example of what I'm trying to create for shading by residuals from a model (complete independence)
#rstats #dataviz #psy6136
NEW in {vcdExtra} 📦 dev version: function color_table() produces colorized versions of flattened n-way frequency tables, (with gt 📦) shading the background of cells according to frequency or residuals from a loglinear model.
[Still a WIP]
Mosaic plot of hair color and eye color, with points ~ observed frequency Code: install.packages("vcdExtra", repos = c('https://friendly.r-universe.dev') HairEye <- margin.table(HairEyeColor, 2:1) mosaic(HairEye, shade = TRUE, legend = FALSE, labeling = labeling_points(scale = 2, seed = 42))
ALT: Mosaic plot of hair color and eye color, with points ~ expected frequency Code: mosaic(HairEye, type = "expected" shade = TRUE, legend = FALSE, labeling = labeling_points(scale = 2, seed = 42))
#rstats #psy6136
NEW in {vcdExtra} 📦 dev version: labeling_points() plots points in the tiles of mosaic diplays showing obs or expected freq, with tiles ~ observed or expected.
Variations in density show pattern of association.
Something I've wanted to do for a long time!
(code in ALT)
#rstats #dataviz #psy6136
Lecture 3 for my course in Categorical Data Analysis, on two-way tables: Tests & visualization
Slides: friendly.github.io/psy6136/lect...
#rstats #datatviz #psy6136
Updating a lecture for my CDA course, I recalled an old paper of mine on why frequency tables naturally area ~ count.
There's a lovely physical analog of gas molecules under pressure that explains everything!
www.tandfonline.com/doi/abs/10.1...
🧵1/4
#rstats #dataviz #psy6136
Lecture 2 for my course in Categorical Data Analysis, on discrete distributions
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
Gave the first lecture for my course in Categorical Data Analysis
Slides: friendly.github.io/psy6136/lect...
#rstats #dataviz #psy6136
Nice separation of greater / less words.
What other way of sorting the words would make any pattern clearer?
