Introduction to R, Class 2: Solutions

Challenge-data

# example1: https://raw.githubusercontent.com/fredhutchio/R_intro/master/extra/clinical.tsv
download.file("https://raw.githubusercontent.com/fredhutchio/R_intro/master/extra/clinical.tsv", "data/clinical.tsv")
clinical_tsv <- read.delim("data/clinical.tsv")
# example2: https://raw.githubusercontent.com/fredhutchio/R_intro/master/extra/clinical.txt
download.file("https://raw.githubusercontent.com/fredhutchio/R_intro/master/extra/clinical.txt", "data/clinical.txt")
clinical_txt <- read.table("data/clinical.txt", header=TRUE)

It’s also possible to read data into R directly using read.csv(URL), but it’s often better practice to download data file for reference.

Challenge-extract

# extract first column, assigning each result to object to show that data structure differs between output
test <- clinical[ , 1]
test2 <- clinical[1]
class(test) # factor (technically, a vector of factors)

## [1] "character"

class(test2) # data.frame

## [1] "data.frame"

Challenge-days

clinical$days_to_death
clinical[6]
clinical[,6]
clinical[[6]]
clinical["days_to_death"] 
clinical[ , "days_to_death"] 
clinical[["days_to_death"]]

Challenge-rows

clinical[1:6, c(3, 6)]

##   age_at_diagnosis days_to_death
## 1            24477           371
## 2            26615           136
## 3            28171          2304
## 4            27154            NA
## 5            29827           146
## 6            23370            NA

Challenge-calculate

range(clinical$cigarettes_per_day, na.rm = TRUE)

## [1]  0.008219178 40.000000000

mean(clinical$cigarettes_per_day, na.rm = TRUE)

## [1] 2.599307

Challenge-not-reported

clinical <- read.csv("../data/clinical.csv", stringsAsFactors = TRUE)
ethnicity <- clinical$ethnicity
levels(ethnicity) # show levels of factor

## [1] "hispanic or latino"     "not hispanic or latino" "not reported"

# rename factors
levels(ethnicity)[3] <- NA
# replace race in data frame
clinical$ethnicity <- ethnicity
# check to see it worked
levels(ethnicity)

## [1] "hispanic or latino"     "not hispanic or latino"

Challenge-remove

Perform a Google search to find additional strategies for renaming missing data. For the search “replace missing data R”, the following result provides a lot of options: https://stackoverflow.com/questions/8161836/how-do-i-replace-na-values-with-zeros-in-an-r-dataframe

Extra exercises

Challenge-disease-race

# show index positions of column names
names(clinical)

##  [1] "primary_diagnosis"           "tumor_stage"                
##  [3] "age_at_diagnosis"            "vital_status"               
##  [5] "morphology"                  "days_to_death"              
##  [7] "state"                       "tissue_or_organ_of_origin"  
##  [9] "days_to_birth"               "site_of_resection_or_biopsy"
## [11] "days_to_last_follow_up"      "cigarettes_per_day"         
## [13] "years_smoked"                "gender"                     
## [15] "year_of_birth"               "race"                       
## [17] "ethnicity"                   "year_of_death"              
## [19] "bcr_patient_barcode"         "disease"

# extract desired object
disease_race <- clinical[6733:6832, c(20, 16)]

Challenge-min-max

min(clinical$days_to_death, na.rm = TRUE)

## [1] 0

max(clinical$days_to_death, na.rm = TRUE)

## [1] 10870

range(clinical$days_to_death, na.rm = TRUE) # does both together

## [1]     0 10870

summary(clinical$days_to_death) # all summary stats for this answer

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
##     0.0   274.0   524.0   878.2  1044.5 10870.0    4645

Challenge-factors

# save column to object
race <- clinical$race 
# recall levels of object
levels(race) 

## [1] "american indian or alaska native"         
## [2] "asian"                                    
## [3] "black or african american"                
## [4] "native hawaiian or other pacific islander"
## [5] "not reported"                             
## [6] "white"

# rename factors
levels(race)[1] <- "American Indian"
levels(race)[2] <- "Asian"
levels(race)[3] <- "black"
levels(race)[4] <- "Pacific Islander"
levels(race)[5] <- NA
# recall levels of object
levels(race) 

## [1] "American Indian"  "Asian"            "black"            "Pacific Islander"
## [5] "white"

# replace race in data frame
clinical$race <- race