UQRUG 19

2021-08-16: UQRUG 19

Participants

• Stéphane Guillou (Library): just here to help!
• Catherine Kim (Library/Biology): working on publishing code with a paper
• Jordan Pennells (ANOVA structure)
• Violeta Berdejo-Espinola (make my code more efficient)
• Omkar Ravindranatha Katagi (Here to learn R)
• … and 8 more UQRUGers!

Shared resources

Links we talked about today:

• Great books / resources to get started:
  • R for Data Science (aka “R4DS”): https://r4ds.had.co.nz/
    • Slack community of R4DS
    • On Twitter: https://twitter.com/R4DScommunity
  • R Cookbook: https://www.cookbook-r.com/
  • RStudio Education: https://education.rstudio.com and https://education.rstudio.com/learn/
  • learnr package (also integrated in the RStudio “tutorial” tab): https://rstudio.github.io/learnr/
  • article on linked points between boxplots: https://datavizpyr.com/how-to-connect-data-points-on-boxplot-with-lines/

Code snippets

Change the order of categorical variable levels (so ggplot2 uses that order instead of the alphabetical order):

library(dplyr)
library(ggplot2)
                          
# relevel factors using forcats package
library(forcats)
f <-  factor(c("a", "b", "c", "d"), levels = c("b", "c", "d", "a"))
fct_relevel(f, "a", "c", "d", "b")

# relevel eye_color of starwars data

starwars %>% 
   mutate(eye_color = fct_relevel(eye_color, "yellow")) %>% 
   ggplot(aes(x = eye_color)) +
   geom_bar()

Use case_when to rate or recode data:

library(dplyr)

starwars
range(starwars$height, na.rm = TRUE)

# just based on height
starwars_rated <-  starwars %>% 
   mutate(height_rating = case_when(height < 70 ~ "very small",
                                    height < 140 ~ "quite small",
                                    height < 200 ~ "medium",
                                    TRUE ~ "tall"))                                       
                                    
# recode based on multiple conditions
range(starwars$mass, na.rm = TRUE) # check range of mass, second vairable to recode by

starwars_rated <-  starwars %>% 
   mutate(height_rating = case_when((height < 70) & (mass < 700)~ "very small and light",
                                    height < 140 ~ "quite small",
                                    height < 200 ~ "medium",
                                    TRUE ~ "tall"))