Factors, dates, and strings

EVR 628- Intro to Environmental Data Science

Author
Affiliation

Juan Carlos Villaseñor-Derbez (JC)

Rosenstiel School of Marine, Atmospheric & Earth Science and Institute for Data Science & Computing

Prompt:

You are given a logbook of fauna sightings during whale watching tours from two vessels operating in the Revillagigedo Archipelago (A.K.A. the Socorro Islands), Mexico. You are about to embark on a tagging expedition, so you want to make sure you get on the right boat. Your goal is therefore to: a) Identify which vessel has the most sightings, and b) Identify the most frequent species sighted.

Exercise 0 - Set up

Post-it up

  1. Add the outline snippet to your snippets.1
  2. Download this csv file to your EVR 628 project folder. Save it into data/raw/.
  3. Start a new script, call it factors_dates_and_strings.R and save it into your scripts/01_processing/ folder
  4. Add a code outline and load the tidyverse and janitor packages at the top of your script

Post-it down

Code 
## SET UP ######################################################################
# Load packages ----------------------------------------------------------------
library(janitor)
library(tidyverse)

Exercise 1 - Dates

Post-it up

  1. Read your data into an object called tour_data <- and clean the column names
  2. Inspect your data. How many columns / rows? What are these columns?
  3. Create a new object called tour_data_clean <- and add a new column called date that contains the date. We will build a large pipeline starting here.
  4. Remove the year, month, and day columns that we no longer need
  5. Place the date column all the way to the left
Code 
# Load data  -------------------------------------------------------------------
tour_data <- read_csv("data/raw/tour_data.csv") |> 
  clean_names()

dim(tour_data)
colnames(tour_data)

## PROCESSING ##################################################################
# Add a date -------------------------------------------------------------------
tour_data_clean <- tour_data |>                    # 3) New object
  mutate(date = make_date(year, month, day)) |>    # 4) Build my date
  select(date, everything(), -c(year, month, day)) # 5) Remove and organize columns
[1] 60  6
[1] "year"       "month"      "day"        "vessel"     "passengers"
[6] "notes"     
The data look like this now:
# A tibble: 6 × 4
  date       vessel passengers notes                                           
  <date>     <chr>       <dbl> <chr>                                           
1 2025-01-01 Condor         21 8 Sea lions; 5 Whales; 4 Whale shark            
2 2025-01-07 Falcon          6 5 whale; 7 Sea lions; 5 Sea turtle; 4 sea turtle
3 2025-01-13 Falcon          8 5 Dolphins                                      
4 2025-01-19 Condor          6 4 Sea turtle; 4 sea turtle                      
5 2025-01-25 Condor         29 2 sea turtle; 8 whale                           
6 2025-02-01 Falcon         19 3 Whale shark; 7 whale

Post-it down

Exercise 2 - Strings

Part 1 - Data contained in a column with strings of text

Post-it up

  1. As demonstrated in class, use separate_longer_delim() to create one row per species mentioned in the data
  2. Standardize the data so all strings appear in lowercase
  3. Make sure they are all singular
Code 
## PROCESSING ##################################################################
# Add a date -------------------------------------------------------------------
tour_data_clean <- tour_data |>                        
  mutate(date = make_date(year, month, day)) |>        
  select(date, everything(), -c(year, month, day)) |>
  separate_longer_delim(cols = notes, delim = ";") |>  # 1) Separate column into new rows
  mutate(notes = str_to_lower(notes),                  # 2) All strings as lower case
         notes = str_remove(notes, "s$"))              # 3) All are singular now

cat("The data look like this now:")
The data look like this now:
Code 
head(tour_data_clean)
# A tibble: 6 × 4
  date       vessel passengers notes           
  <date>     <chr>       <dbl> <chr>           
1 2025-01-01 Condor         21 "8 sea lion"    
2 2025-01-01 Condor         21 " 5 whale"      
3 2025-01-01 Condor         21 " 4 whale shark"
4 2025-01-07 Falcon          6 "5 whale"       
5 2025-01-07 Falcon          6 " 7 sea lion"   
6 2025-01-07 Falcon          6 " 5 sea turtle"

Post-it down

Part 2 - Oh no! Data are STILL contained in a column with strings of text!

Post-it up

  1. Look at the documentation for str_extract()
  2. Run the first two examples at the bottom directly in your console. What seems to be going on?

pause to discuss

  1. Using the str_extract() function, build a new column called fauna that contains the species observed2
  2. Print your data to the console. Does fauna look right?
  3. Using the str_extract() function, build a new column called n that contains the number of organisms observed (make sure it’s numeric)
  4. Remove the notes column
Code 
## PROCESSING ##################################################################
# Add a date -------------------------------------------------------------------
tour_data_clean <- tour_data |>                        
  mutate(date = make_date(year, month, day)) |>        
  select(date, everything(), -c(year, month, day)) |>
  separate_longer_delim(cols = notes, delim = ";") |>
  mutate(notes = str_to_lower(notes),
         notes = str_remove(notes, "s$"),
         fauna = str_extract(string = notes, pattern = "[a-z ]+$"), # 3) Extract fauna
         fauna = str_squish(fauna),                                 # 4) Remove white space
         n = str_extract(string = notes, pattern = "[:digit:]+"),   # 5) Extract n
         n = as.numeric(n)) |>                                            # and make sure it's numeric
  select(-notes)                                                      # 6) Remove notes

cat("The data look like this now:")
The data look like this now:
Code 
head(tour_data_clean)
# A tibble: 6 × 5
  date       vessel passengers fauna           n
  <date>     <chr>       <dbl> <chr>       <dbl>
1 2025-01-01 Condor         21 sea lion        8
2 2025-01-01 Condor         21 whale           5
3 2025-01-01 Condor         21 whale shark     4
4 2025-01-07 Falcon          6 whale           5
5 2025-01-07 Falcon          6 sea lion        7
6 2025-01-07 Falcon          6 sea turtle      5

Post-it down

Exercise 3 - Visualization (a bit of everything)

Post-it up

  1. Build a new data frame called tour_data_viz that contains the following columns3:
  • vessel
  • fauna
  • n_per_trip
  1. Recreate the plot below

Post-it down

Code 
## Build tour_data_viz ---------------------------------------------------------
tour_data_viz <- tour_data_clean |> 
  group_by(vessel) |> 
  mutate(effort = n_distinct(date)) |> 
  group_by(vessel, fauna) |> 
  summarize(n_per_trip = sum(n / effort),
            .groups = "drop")

cat("The data look like this now:")
The data look like this now:
Code 
head(tour_data_viz)
# A tibble: 6 × 3
  vessel fauna       n_per_trip
  <chr>  <chr>            <dbl>
1 Condor dolphin           0.25
2 Condor sea lion          1.56
3 Condor sea turtle        2.06
4 Condor whale             3.25
5 Condor whale shark       1   
6 Falcon dolphin           1.43
Code 
## VISUALIZE ###################################################################
ggplot(data = tour_data_viz,
       mapping = aes(x = str_to_sentence(fauna), # This is the only str-relevant change here
                     y = n_per_trip)) +
  geom_col() +
  facet_wrap(~vessel, ncol = 1) +
  labs(title = "Sightings by vessel",
       x = "Fauna",
       y = "N") +
  theme_minimal(base_size = 14)

Footnotes

  1. Hint: Tools > Edit Code Snippets… Remember to use tabs, not spaces↩︎

  2. Hint: Look at the documentation for ?regex↩︎

  3. Hint, you will need two group_bys, one mutate, and one summarize↩︎