Homework 7

  1. At the top of a new script, include code for loading basecase, dplyr, and ggplot2. Also include code for importing the nhanes_l.csv data file. Run your code.

    library(baseverse)
library(dplyr)
library(ggplot2)
nhanes<-read.csv('nhanes_l.csv')

  2. Use base_match() to create a labeled version of the original arthritis variable (mcq160a). Values of 1 represent people with a history of arthritis and values of 0 represent people without a history of arthritis. The NHANES documentation for mcq106a is available here. List the no group first. Check your work and verify that the groups are listed in the desired order.

    nhanes$arthritis<-base_match(nhanes$mcq160a,'No'=2,'Arthritis'=1)
table(nhanes$mcq160a,useNA='always')
#|  
#|     1    2    9 <NA> 
#|  2532 5258   17  346
table(nhanes$arthritis,useNA='always')
#|  
#|         No Arthritis      <NA> 
#|       5258      2532       363

  3. Use base_when() and the original glucose variable (lbxglu) to define a categorical variable for glucose, using the following definitions:

    • lbxglu < 100 → Normal
    • 100 ≤ lbxglu < 126 → Prediabetes
    • lbxglu ≥ 126 → Diabetes

    List the groups in the above order. Check your work and verify that the groups are listed in table() in the desired order.

    nhanes$diabetes<-base_when(
   'Normal' = nhanes$lbxglu<100,
   'Prediabetes' = nhanes$lbxglu>=100 & nhanes$lbxglu<126,
   'Diabetes' = nhanes$lbxglu>=126
)
sum(nhanes$lbxglu<100,na.rm=TRUE)
#|  [1] 1499
sum(nhanes$lbxglu>=100 & nhanes$lbxglu<126,na.rm=TRUE)
#|  [1] 1401
sum(nhanes$lbxglu>=126,na.rm=TRUE)
#|  [1] 429
table(nhanes$diabetes,useNA='always')
#|  
#|       Normal Prediabetes    Diabetes        <NA> 
#|         1499        1401         429        4824

  4. Use ggplot2 to create a violin plot for the new categorical variable (above) versus the continuous glucose variable (lbxglu). Exclude people missing lbxglu. You can interpret your plot as a proof or check of how the categorical variable was defined.

    ggplot(subset(nhanes,!is.na(lbxglu)),aes(x=diabetes,y=lbxglu))+
   geom_violin()+
  labs(x='Diabetes status',y='Fasting glucose (mg/dL)')