This lab is to test the main concepts and functions you need to make data tidy. You’ll work a generically named dummy dataset called whales.
The “data” were collected as follows: observers are asked for certain information about specific indicents they witnessed of ships striking whales and that information is compiled by whale type. The observers were asked to provide: type of whale, date of event (m/d/yr), outcome of event, approximate length of whale in feet, ocean in which event occurred.
Sometimes an observer could not provide all of that information, and missing data is represented as blanks between commas—look at the dataset to see. An observer can possibly give information about more than one event.
Such ways of organizing information are not that uncommon, and you should expect to see many varied formats for raw data—which you will need to grapple with before analysis.
If the instructions say to create an object or variable with a certain name, do it exactly as written in the prompt.
library(tidyverse)
whales <- read_csv("https://raw.githubusercontent.com/idc9/stor390/master/data/whales.csv")
str(whales)## Classes 'tbl_df', 'tbl' and 'data.frame': 31 obs. of 9 variables:
## $ observer : int 1 2 3 4 5 6 7 8 9 10 ...
## $ blue : chr "1/20/15, death, , Indian" NA NA NA ...
## $ humpback : chr NA "8/12/15, death, 50, atlantic" NA "3/4/12, death, 56, pacific" ...
## $ southern_right: chr NA NA "7/14/13, injury, 47, pacific" NA ...
## $ sei : chr "8/9/11, injury, , indian" NA NA NA ...
## $ fin : chr NA "8/2/13, death, 76, arctic" NA NA ...
## $ killer_whale : chr NA NA NA NA ...
## $ bowhead : chr NA "6/24/13, injury, 30, artic" NA NA ...
## $ grey : chr NA NA NA "5/24/16, death, , pacific" ...
## - attr(*, "spec")=List of 2
## ..$ cols :List of 9
## .. ..$ observer : list()
## .. .. ..- attr(*, "class")= chr "collector_integer" "collector"
## .. ..$ blue : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ humpback : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ southern_right: list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ sei : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ fin : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ killer_whale : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ bowhead : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## .. ..$ grey : list()
## .. .. ..- attr(*, "class")= chr "collector_character" "collector"
## ..$ default: list()
## .. ..- attr(*, "class")= chr "collector_guess" "collector"
## ..- attr(*, "class")= chr "col_spec"Since the meaning of tidy depends on your goals, assume you are someone interested in describing ship strike events on whales based on this dataset. Your unit of observation is an event.
Use whatever functions you deem necessary to answer the question: Does whales meet the three criteria for tidy data given in the lecture? Put the code you wrote in one chunk then type TRUE or FALSE in another code chunk.
#Create a data frame that has one row per observer, per species and one single variable of all the information collected. This is an example of a key-value pair.
Label your key variable species and your value variable info.
Your new object should write over the original whales data frame.
#Create a data frame that includes only events for which there is information—writing over the whales object again.
#hint: is.na
Create a data frame with one variable per type of information, one piece of information per cell. Some cells might be empty.
Again replace the old whales with the new.
Your new data frame should have six variables: observer, species, date, outcome, size, ocean.
#Do what you need to do to make the variables be of the following types (listed in order): integer, character, datetime (Y-M-D), character, integer, character.
All character and factor variables should be entirely in lower case letters.
Using an ifelse statement or other method, replace blanks in your data with NA.
Again save your result as whales.
#hint: parse_datetime
Print a summary table with: 1) number ship strikes by species, 2) average whale size by species, omitting NA values in the calculation.
Print here means you do not need to save the result.
#As in the lecture, use unite to check the dataset has only one observation per observer and species. You do not need to save the result, just print a summary as in the lecture.
#Return the dataset to its original configuration: One row per observer, one column per species, one cell for all information with individual variables separated by commas.
Don’t worry about the NA values you replaced blanks with, and don’t worry about the change in date format or any other changes in variable format.
Do put the information back in the same order in which it came.
Save over the old whales object with the new.
#