ggplot 04 Collective geoms скачать в хорошем качестве

ggplot 04 Collective geoms

In this lecture, we move past "what geom should I use?" and focus on the concept that quietly determines whether your ggplot is correct: grouping—especially for collective geoms like lines, paths, polygons, and statistical summaries. You will learn how ggplot2 decides which observations belong together, why "wrong grouping" produces classic failure modes (including the sawtooth line artifact), and how grouping changes not only what gets drawn but also how statistical layers compute results. What you’ll learn (practical takeaways) Individual vs. collective geoms and why this distinction matters for interpretation. How the group aesthetic is chosen by default (and when you must override it). Three common scenarios where default grouping fails - plus reliable fixes. How to build and interpret "spaghetti plots" for longitudinal data (many subjects, one visual style). How to overlay trajectories on discrete x scales (e.g., boxplots and lines) by overriding grouping. Why grouping affects stats (e.g., why you accidentally get one smoother per subject—and how to avoid it). How aesthetics map onto collective objects, including the “first value” principle for lines/paths. Dataset used We work through examples using the Oxboys longitudinal dataset (26 boys measured across nine occasions) to make grouping behavior obvious and memorable. Resources ggplot2 book (online): https://ggplot2-book.org/ Book source repository: https://github.com/hadley/ggplot2-book/ If you found this helpful, please like and subscribe. #rstats #ggplot2 #datavisualization #datascience #statistics #tidyverse