Prerequisites

Install the following packages:

  • tidyr
  • dplyr
  • ggplot2

Load the libraries with R:

library(tidyr)
## Warning: 程辑包'tidyr'是用R版本4.3.2 来建造的
library(dplyr)
## Warning: 程辑包'dplyr'是用R版本4.3.2 来建造的
## 
## 载入程辑包：'dplyr'
## The following objects are masked from 'package:stats':
## 
##     filter, lag
## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union
library(ggplot2)
## Warning: 程辑包'ggplot2'是用R版本4.3.2 来建造的

Section Example: weather data from BaoAn airport

Environment data, expecially raw observations, are really messy. Here we take a look at hourly weather data measured at BaoAn International Airport during the past 10 years. The data set is from NOAA Integrated Surface Dataset.

Download the file 2281305.zip, where the number 2281305 is the site ID.

Extract the zip file, you should see a file named 2281305.csv.

Open the file, you see many columns and even more lines, with no clear meaning and definitions. This is very often you would encounter when work with environmental data. As Hadley Wickham puts it, “Tidy datasets are all alike, but every messy dataset is messy in its own way.”

The notes below are modified from the excellent Dataframe Manipulation and freely available on the Software Carpentry website.

Loading a .csv file

drawing

Figure source

Charles David Keeling directed a program to measure the concentrations of CO2 in the atmosphere that continued without interruption from the late 1950s through the present. This program, operated out of Scripps Institution of Oceanography, is responsible for the Mauna Loa record, which is almost certainly the best-known icon illustrating the impact of humanity on the planet as a whole. Learn more about Keeling Curve Lessons.

We will start with monthly CO2 data measured from Mauna Loa.

To begin with,

  • Download the co2_mm_mlo.csv file from here

  • Save the file to your working directory.

  • Take a look at the file, where the column co2 means monthly average CO2 in the unit of ppm (part per million, 10-6), quality column is the quality flag of the observation, 0 means the data point does not meet the quality control so that should be discarded, 1 means the data point is usable.

Now let’s load this file via the following:

Keeling_Data <- read.csv(file = "co2_mm_mlo.csv", header = T)

The file is loaded into the dataset Keeling_Data. Here we use the option header = T so that the first line of co2_mm_mlo.csv is also loaded as names of the variables. You can try to turn this off by setting header = F, and check what happens.

Check the names of columns:

colnames(Keeling_Data)
## [1] "year"         "month"        "decimal_date" "co2"          "quality"

If column names are not specified, e.g., using headers = FALSE in a read.csv() function, R assigns default names V1, V2, …, Vn.

Check the head (first 6 lines) of the dataset using head() function

head(Keeling_Data)
##   year month decimal_date    co2 quality
## 1 1958     3     1958.203 315.70       1
## 2 1958     4     1958.288 317.45       1
## 3 1958     5     1958.370 317.51       1
## 4 1958     6     1958.455 -99.99       0
## 5 1958     7     1958.537 315.86       1
## 6 1958     8     1958.622 314.93       1

Or the end of (last 6 lines) with tail():

tail(Keeling_Data)
##     year month decimal_date    co2 quality
## 786 2023     8     2023.625 419.68       1
## 787 2023     9     2023.708 418.51       1
## 788 2023    10     2023.792 418.82       1
## 789 2023    11     2023.875 420.46       1
## 790 2023    12     2023.958 421.86       1
## 791 2024     1     2024.042 422.80       1

The read.table() function is used for reading in tabular data stored in a text file where the columns of data are separated by punctuation characters such as .csv files (csv = comma-separated values). Tabular files are the most common file type you would encounter in your future study, as it’s easy to use and share.

Tabs and commas are the most common punctuation characters used to separate or delimit data points in .csv files. For convenience R provides 2 other versions of read.table(). These are: read.csv() for files where the data are separated with commas and read.delim() for files where the data are separated with tabs. Of these three functions, read.csv() is the most commonly used. If needed it is possible to override the default delimiting punctuation marks for both read.csv() and read.delim().

We can begin exploring our data set right away, pulling out columns by specifying them using the $ operator:

Keeling_Data$co2
##   [1] 315.70 317.45 317.51 -99.99 315.86 314.93 313.20 -99.99 313.33 314.67 315.58 316.48 316.65 317.72 318.29 318.15 316.54 314.80 313.84 313.33 314.81 315.58 316.43 316.98 317.58 319.03 320.04 319.59 318.18 315.90
##  [31] 314.17 313.83 315.00 316.19 316.89 317.70 318.54 319.48 320.58 319.77 318.57 316.79 314.99 315.31 316.10 317.01 317.94 318.55 319.68 320.57 321.02 320.62 319.61 317.40 316.25 315.42 316.69 317.70 318.74 319.07
##  [61] 319.86 321.38 322.25 321.48 319.74 317.77 316.21 315.99 317.07 318.35 319.57 -99.99 -99.99 -99.99 322.26 321.89 320.44 318.69 316.70 316.87 317.68 318.71 319.44 320.44 320.89 322.14 322.17 321.87 321.21 318.87
##  [91] 317.81 317.30 318.87 319.42 320.62 321.60 322.39 323.70 324.08 323.75 322.38 320.36 318.64 318.10 319.78 321.03 322.33 322.50 323.04 324.42 325.00 324.09 322.54 320.92 319.25 319.39 320.73 321.96 322.57 323.15
## [121] 323.89 325.02 325.57 325.36 324.14 322.11 320.33 320.25 321.32 322.89 324.00 324.42 325.63 326.66 327.38 326.71 325.88 323.66 322.38 321.78 322.86 324.12 325.06 325.98 326.93 328.13 328.08 327.67 326.34 324.69
## [151] 323.10 323.06 324.01 325.13 326.17 326.68 327.17 327.79 328.93 328.57 327.36 325.43 323.36 323.56 324.80 326.01 326.77 327.63 327.75 329.72 330.07 329.09 328.04 326.32 324.84 325.20 326.50 327.55 328.55 329.56
## [181] 330.30 331.50 332.48 332.07 330.87 329.31 327.51 327.18 328.16 328.64 329.35 330.71 331.48 332.65 333.18 332.20 331.07 329.15 327.33 327.28 328.31 329.58 330.73 331.46 331.94 333.11 333.95 333.42 331.97 329.95
## [211] 328.50 328.36 329.38 -99.99 331.56 332.74 333.36 334.74 334.72 333.98 333.08 330.68 328.96 328.72 330.16 331.62 332.68 333.17 334.96 336.14 336.93 336.17 334.89 332.56 331.29 331.28 332.46 333.60 334.94 335.26
## [241] 336.66 337.69 338.02 338.01 336.50 334.42 332.36 332.45 333.76 334.91 336.14 336.69 338.27 338.82 339.24 339.26 337.54 335.72 333.97 334.24 335.32 336.81 337.90 338.34 340.07 340.93 341.45 341.36 339.45 337.67
## [271] 336.25 336.14 337.30 338.29 339.29 340.55 341.63 342.60 343.04 342.54 340.82 338.48 336.95 337.05 338.58 339.91 340.93 341.76 342.78 343.96 344.77 343.88 342.42 340.24 338.38 338.41 339.44 340.78 341.57 342.79
## [301] 343.37 345.39 346.14 345.76 344.32 342.51 340.46 340.53 341.79 343.20 344.21 344.92 345.68 -99.99 347.78 347.16 345.79 343.74 341.59 341.86 343.31 345.00 345.48 346.42 347.91 348.66 349.28 348.65 346.90 345.26
## [331] 343.47 343.35 344.73 346.12 346.78 347.48 348.25 349.86 350.52 349.98 348.25 346.17 345.48 344.82 346.22 347.49 348.73 348.92 349.81 351.40 352.15 351.58 350.21 348.20 346.66 346.72 348.08 349.28 350.51 351.70
## [361] 352.50 353.67 354.35 353.88 352.80 350.49 348.97 349.37 350.42 351.62 353.07 353.43 354.08 355.72 355.95 355.44 354.05 351.84 350.09 350.33 351.55 352.91 353.86 355.10 355.75 356.38 357.38 356.39 354.89 353.06
## [391] 351.38 351.69 353.14 354.41 354.93 355.82 357.33 358.77 359.23 358.23 356.30 353.97 352.34 352.43 353.89 355.21 356.34 357.21 357.97 359.22 359.71 359.44 357.15 354.99 353.01 353.41 354.42 355.68 357.10 357.42
## [421] 358.59 359.39 360.30 359.64 357.45 355.76 354.14 354.23 355.53 357.03 358.36 359.04 360.11 361.36 361.78 360.94 359.51 357.59 355.86 356.21 357.65 359.10 360.04 361.00 361.98 363.44 363.83 363.33 361.78 359.33
## [451] 358.32 358.14 359.61 360.82 362.20 363.36 364.28 364.69 365.25 365.06 363.69 361.55 359.69 359.72 361.04 362.39 363.24 364.21 364.65 366.48 366.77 365.73 364.46 362.40 360.44 360.98 362.65 364.51 365.39 366.10
## [481] 367.36 368.79 369.56 369.13 367.98 366.10 364.16 364.54 365.67 367.30 368.35 369.28 369.84 371.15 371.12 370.46 369.61 367.06 364.95 365.52 366.88 368.26 369.45 369.71 370.75 371.98 371.75 371.87 370.02 368.27
## [511] 367.15 367.18 368.53 369.83 370.76 371.69 372.63 373.55 374.03 373.40 371.68 369.78 368.34 368.61 369.94 371.42 372.70 373.37 374.30 375.19 375.93 375.69 374.16 372.03 370.92 370.73 372.43 373.98 375.07 375.82
## [541] 376.64 377.92 378.78 378.46 376.88 374.57 373.34 373.31 374.84 376.17 377.17 378.05 379.06 380.54 380.80 379.87 377.65 376.17 374.43 374.63 376.33 377.68 378.63 379.91 380.95 382.48 382.64 382.40 380.93 378.93
## [571] 376.89 377.19 378.54 380.31 381.58 382.40 382.86 384.80 385.22 384.24 382.65 380.60 379.04 379.33 380.35 382.02 383.10 384.12 384.81 386.73 386.78 386.33 384.73 382.24 381.20 381.37 382.70 384.19 385.78 386.06
## [601] 386.28 387.34 388.78 387.99 386.61 384.32 383.41 383.21 384.41 385.79 387.17 387.70 389.04 389.76 390.36 389.70 388.24 386.29 384.95 384.64 386.23 387.63 388.91 390.41 391.37 392.67 393.21 392.38 390.41 388.54
## [631] 387.03 387.43 388.87 389.99 391.50 392.05 392.80 393.44 394.41 393.95 392.72 390.33 389.28 389.19 390.48 392.06 393.31 394.04 394.59 396.38 396.93 395.91 394.56 392.59 391.32 391.27 393.20 394.57 395.78 397.03
## [661] 397.66 398.64 400.02 398.81 397.51 395.39 393.72 393.90 395.36 397.03 398.04 398.27 399.91 401.51 401.96 401.43 399.27 397.18 395.54 396.16 397.40 399.08 400.18 400.55 401.74 403.34 404.15 402.97 401.46 399.11
## [691] 397.82 398.49 400.27 402.06 402.73 404.25 405.06 407.60 407.90 406.99 404.59 402.45 401.23 401.79 403.72 404.64 406.36 406.66 407.54 409.22 409.89 409.08 407.33 405.32 403.57 403.82 405.31 407.00 408.15 408.52
## [721] 409.59 410.45 411.44 410.99 408.90 407.16 405.71 406.19 408.21 409.27 411.03 411.96 412.18 413.54 414.86 414.15 411.96 410.17 408.76 408.74 410.47 411.97 413.59 414.32 414.71 416.42 417.28 416.58 414.58 412.75
## [751] 411.49 411.48 413.10 414.23 415.49 416.72 417.61 419.01 419.09 418.92 416.90 414.42 413.26 413.90 414.97 416.67 418.13 419.24 418.76 420.19 420.97 420.94 418.85 417.15 415.91 415.74 417.47 418.99 419.48 420.30
## [781] 420.98 423.35 424.00 423.68 421.83 419.68 418.51 418.82 420.46 421.86 422.80

Let’s do some simple statistical checks with Keeling_Data$co2:

# Min
min(Keeling_Data$co2)
## [1] -99.99
# Max
max(Keeling_Data$co2)
## [1] 424
# Range
range(Keeling_Data$co2)
## [1] -99.99 424.00
# Mean
mean(Keeling_Data$co2)
## [1] 355.0285
# Median
median(Keeling_Data$co2)
## [1] 354.93
# Show quantiles
summary(Keeling_Data$co2)
##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  -99.99  330.01  354.93  355.03  384.28  424.00

You will find there are some -99.99 values, which are missing values. We will get back to this later.

You can use [] to extract elements of a vector by specifying their corresponding index.

Important: Index in R starts from 1, not 0. For example:

Keeling_Data$co2[1:10]
##  [1] 315.70 317.45 317.51 -99.99 315.86 314.93 313.20 -99.99 313.33 314.67
Keeling_Data$year[200:210]
##  [1] 1974 1974 1974 1975 1975 1975 1975 1975 1975 1975 1975
Keeling_Data$co2[0]
## numeric(0)

Tidy data

In the rest of the section, we will learn a consistent way to organize the data in R, called tidy data. Getting the data into this format requires some upfront work, but that work pays off in the long term. Once you have tidy data and the tools provided by the tidyr, dplyr and ggplot2 packages, you will spend much less time munging/wrangling data from one representation to another, allowing you to spend more time on the analytic questions at hand.

Tibble

For now, we will use tibble instead of R’s traditional data.frame. Tibble is a data frame, but they tweak some older behaviors to make life a little easier.

Let’s use Keeling_Data again:

Keeling_Data <- read.csv(file = "co2_mm_mlo.csv", header = T)
head(Keeling_Data)
##   year month decimal_date    co2 quality
## 1 1958     3     1958.203 315.70       1
## 2 1958     4     1958.288 317.45       1
## 3 1958     5     1958.370 317.51       1
## 4 1958     6     1958.455 -99.99       0
## 5 1958     7     1958.537 315.86       1
## 6 1958     8     1958.622 314.93       1

You can coerce a data.frame to a tibble using the as_tibble() function:

Keeling_Data_tbl <- as_tibble(Keeling_Data)
Keeling_Data_tbl
## # A tibble: 791 × 5
##     year month decimal_date   co2 quality
##    <int> <int>        <dbl> <dbl>   <int>
##  1  1958     3        1958.  316.       1
##  2  1958     4        1958.  317.       1
##  3  1958     5        1958.  318.       1
##  4  1958     6        1958. -100.       0
##  5  1958     7        1959.  316.       1
##  6  1958     8        1959.  315.       1
##  7  1958     9        1959.  313.       1
##  8  1958    10        1959. -100.       0
##  9  1958    11        1959.  313.       1
## 10  1958    12        1959.  315.       1
## # ℹ 781 more rows

Tidy dataset

There are three inter-related rules which make a dataset tidy:

  • Each variable must have its own column;
  • Each observation must have its own row;
  • Each value must have its own cell;

These three rules are interrelated because it’s impossible to only satisfy two of the three.

That interrelationship leads to an even simpler set of practical instructions:

  • Put each dataset in a tibble
  • Put each variable in a column

Why ensure that your data is tidy? There are two main advantages:

  • There’s a general advantage to picking one consistent way of storing data. If you have a consistent data structure, it’s easier to learn the tools that work with it because they have an underlying uniformity. If you ensure that your data is tidy, you’ll spend less time fighting with the tools and more time working on your analysis.

  • There’s a specific advantage to placing variables in columns because it allows R’s vectorized nature to shine. That makes transforming tidy data feel particularly natural. tidyr, dplyr, and ggplot2 are designed to work with tidy data.

The dplyr package

The dplyr package provides a number of very useful functions for manipulating dataframes in a way that will reduce the self-repetition, reduce the probability of making errors, and probably even save you some typing. As an added bonus, you might even find the dplyr grammar easier to read.

Here we’re going to cover commonly used functions as well as using pipes %>% to combine them.

Using select()

Use the select() function to keep only the variables (columns) you select.

dplyr::select(Keeling_Data_tbl, year, co2, quality)
## # A tibble: 791 × 3
##     year   co2 quality
##    <int> <dbl>   <int>
##  1  1958  316.       1
##  2  1958  317.       1
##  3  1958  318.       1
##  4  1958 -100.       0
##  5  1958  316.       1
##  6  1958  315.       1
##  7  1958  313.       1
##  8  1958 -100.       0
##  9  1958  313.       1
## 10  1958  315.       1
## # ℹ 781 more rows
dplyr::select(Keeling_Data_tbl, co2)
## # A tibble: 791 × 1
##      co2
##    <dbl>
##  1  316.
##  2  317.
##  3  318.
##  4 -100.
##  5  316.
##  6  315.
##  7  313.
##  8 -100.
##  9  313.
## 10  315.
## # ℹ 781 more rows

The pipe symbol %>%

Above we used ‘normal’ grammar, but the strengths of dplyr and tidyr lie in combining several functions using pipes. Since the pipes grammar is unlike anything we’ve seen in R before, let’s repeat what we’ve done above using pipes.

x %>% f(y) is the same as f(x, y)

Keeling_Data_tbl %>% 
  dplyr::select(year, co2, quality)
## # A tibble: 791 × 3
##     year   co2 quality
##    <int> <dbl>   <int>
##  1  1958  316.       1
##  2  1958  317.       1
##  3  1958  318.       1
##  4  1958 -100.       0
##  5  1958  316.       1
##  6  1958  315.       1
##  7  1958  313.       1
##  8  1958 -100.       0
##  9  1958  313.       1
## 10  1958  315.       1
## # ℹ 781 more rows

The above lines mean we first call the Keeling_Data_tbl tibble and pass it on, using the pipe symbol %>%, to the next step, which is the select() function. In this case, we don’t specify which data object we use in the select() function since in gets that from the previous pipe. The select() function then takes what it gets from the pipe, in this case the Keeling_Data_tbl tibble, as its first argument. By using pipe, we can take output of the previous step as input for the next one, so that we can avoid defining and calling unnecessary temporary variables. You will start to see the power of pipe later.

Using filter()

Use filter() to get values (rows):

Keeling_Data_tbl %>% 
  filter(year == 2000)
## # A tibble: 12 × 5
##     year month decimal_date   co2 quality
##    <int> <int>        <dbl> <dbl>   <int>
##  1  2000     1        2000.  369.       1
##  2  2000     2        2000.  370.       1
##  3  2000     3        2000.  371.       1
##  4  2000     4        2000.  372.       1
##  5  2000     5        2000.  372.       1
##  6  2000     6        2000.  372.       1
##  7  2000     7        2001.  370.       1
##  8  2000     8        2001.  368.       1
##  9  2000     9        2001.  367.       1
## 10  2000    10        2001.  367.       1
## 11  2000    11        2001.  369.       1
## 12  2000    12        2001.  370.       1

If we now want to move forward with the above tibble, but only with quality == 1 , we can combine select() and filter() functions:

Keeling_Data_tbl %>% 
  dplyr::select(year, co2, quality) %>% 
  filter(quality == 1)
## # A tibble: 784 × 3
##     year   co2 quality
##    <int> <dbl>   <int>
##  1  1958  316.       1
##  2  1958  317.       1
##  3  1958  318.       1
##  4  1958  316.       1
##  5  1958  315.       1
##  6  1958  313.       1
##  7  1958  313.       1
##  8  1958  315.       1
##  9  1959  316.       1
## 10  1959  316.       1
## # ℹ 774 more rows

You see here we have used the pipe twice, and the scripts become really clean and easy to follow.

Using group_by() and summarize()

Now try to group monthly data using the group_by() function, notice how the output tibble changes:

Keeling_Data_tbl %>% 
  dplyr::select(year, month, co2, quality) %>% 
  filter(quality == 1) %>% 
  group_by(month)
## # A tibble: 784 × 4
## # Groups:   month [12]
##     year month   co2 quality
##    <int> <int> <dbl>   <int>
##  1  1958     3  316.       1
##  2  1958     4  317.       1
##  3  1958     5  318.       1
##  4  1958     7  316.       1
##  5  1958     8  315.       1
##  6  1958     9  313.       1
##  7  1958    11  313.       1
##  8  1958    12  315.       1
##  9  1959     1  316.       1
## 10  1959     2  316.       1
## # ℹ 774 more rows

The group_by() function is much more exciting in conjunction with the summarize() function. This will allow us to create new variable(s) by using functions that repeat for each of the continent-specific data frames. That is to say, using the group_by() function, we split our original dataframe into multiple pieces, then we can run functions (e.g., mean() or sd()) within summarize().

Keeling_Data_tbl %>% 
  dplyr::select(year, month, co2, quality) %>% 
  filter(quality == 1) %>% 
  group_by(month) %>% 
  summarize(monthly_mean = mean(co2))
## # A tibble: 12 × 2
##    month monthly_mean
##    <int>        <dbl>
##  1     1         360.
##  2     2         360.
##  3     3         360.
##  4     4         362.
##  5     5         361.
##  6     6         362.
##  7     7         359.
##  8     8         357.
##  9     9         356.
## 10    10         356.
## 11    11         357.
## 12    12         359.

Here we create a new variable (column) monthly_mean, and append it to the groups (month in this case). Now, we get a so-called monthly climatology.

You can also use arrange() and desc() to sort the data:

Keeling_Data_tbl %>% 
  dplyr::select(year, month, co2, quality) %>% 
  filter(quality == 1) %>% 
  group_by(month) %>% 
  summarize(monthly_mean = mean(co2)) %>% 
  arrange(monthly_mean)
## # A tibble: 12 × 2
##    month monthly_mean
##    <int>        <dbl>
##  1     9         356.
##  2    10         356.
##  3    11         357.
##  4     8         357.
##  5    12         359.
##  6     7         359.
##  7     1         360.
##  8     2         360.
##  9     3         360.
## 10     5         361.
## 11     6         362.
## 12     4         362.
Keeling_Data_tbl %>% 
  dplyr::select(year, month, co2, quality) %>% 
  filter(quality == 1) %>% 
  group_by(month) %>% 
  summarize(monthly_mean = mean(co2)) %>% 
  arrange(desc(monthly_mean))
## # A tibble: 12 × 2
##    month monthly_mean
##    <int>        <dbl>
##  1     4         362.
##  2     6         362.
##  3     5         361.
##  4     3         360.
##  5     2         360.
##  6     1         360.
##  7     7         359.
##  8    12         359.
##  9     8         357.
## 10    11         357.
## 11    10         356.
## 12     9         356.

Let’s add more statistics to the monthly climatology:

Keeling_Data_tbl %>% 
  dplyr::select(year, month, co2, quality) %>% 
  filter(quality == 1) %>% 
  group_by(month) %>% 
  summarize(monthly_mean = mean(co2), monthly_sd = sd(co2),
            monthly_min = min(co2), monthly_max = max(co2),
            monthly_se = sd(co2)/sqrt(n()))
## # A tibble: 12 × 6
##    month monthly_mean monthly_sd monthly_min monthly_max monthly_se
##    <int>        <dbl>      <dbl>       <dbl>       <dbl>      <dbl>
##  1     1         360.       32.0        316.        423.       3.94
##  2     2         360.       31.1        316.        420.       3.89
##  3     3         360.       31.3        316.        421.       3.89
##  4     4         362.       31.7        317.        423.       3.96
##  5     5         361.       31.6        318.        424        3.89
##  6     6         362.       31.4        318.        424.       3.89
##  7     7         359.       31.4        316.        422.       3.86
##  8     8         357.       31.3        315.        420.       3.86
##  9     9         356.       31.4        313.        419.       3.87
## 10    10         356.       31.4        313.        419.       3.89
## 11    11         357.       31.8        313.        420.       3.91
## 12    12         359.       31.9        315.        422.       3.96

Here we call the n() to get the size of a vector.

Using mutate()

We can also create new variables (columns) using the mutate() function. Here we create a new column co2_ppb by simply scaling co2 by a factor of 1000.

Keeling_Data_tbl %>% 
  mutate(co2_ppb = co2 * 1e3)
## # A tibble: 791 × 6
##     year month decimal_date   co2 quality co2_ppb
##    <int> <int>        <dbl> <dbl>   <int>   <dbl>
##  1  1958     3        1958.  316.       1  315700
##  2  1958     4        1958.  317.       1  317450
##  3  1958     5        1958.  318.       1  317510
##  4  1958     6        1958. -100.       0  -99990
##  5  1958     7        1959.  316.       1  315860
##  6  1958     8        1959.  315.       1  314930
##  7  1958     9        1959.  313.       1  313200
##  8  1958    10        1959. -100.       0  -99990
##  9  1958    11        1959.  313.       1  313330
## 10  1958    12        1959.  315.       1  314670
## # ℹ 781 more rows

When creating new variables, we can hook this with a logical condition. A simple combination of mutate() and ifelse() facilitates filtering right where it is needed: in the moment of creating something new. This easy-to-read statement is a fast and powerful way of discarding certain data or for updating values depending on this given condition.

Let’s create a new variable co2_new, it is equal to co2 when quality==1, otherwise it’s NA:

Keeling_Data_tbl %>% 
  mutate(co2_ppb = co2 * 1e3, co2_new = ifelse(quality==1, co2, NA))
## # A tibble: 791 × 7
##     year month decimal_date   co2 quality co2_ppb co2_new
##    <int> <int>        <dbl> <dbl>   <int>   <dbl>   <dbl>
##  1  1958     3        1958.  316.       1  315700    316.
##  2  1958     4        1958.  317.       1  317450    317.
##  3  1958     5        1958.  318.       1  317510    318.
##  4  1958     6        1958. -100.       0  -99990     NA 
##  5  1958     7        1959.  316.       1  315860    316.
##  6  1958     8        1959.  315.       1  314930    315.
##  7  1958     9        1959.  313.       1  313200    313.
##  8  1958    10        1959. -100.       0  -99990     NA 
##  9  1958    11        1959.  313.       1  313330    313.
## 10  1958    12        1959.  315.       1  314670    315.
## # ℹ 781 more rows

Combining dplyr and ggplot2

Just as we used %>% to pipe data along a chain of dplyr functions we can use it to pass data to ggplot(). Because %>% replaces the first argument in a function we don’t need to specify the data = argument in the ggplot() function.

By combining dplyr and ggplot2 functions, we can make figures without creating any new variables or modifying the data.

Keeling_Data_tbl %>% 
  mutate(co2_new = ifelse(quality==1, co2, NA)) %>% 
  # Make the plot
  ggplot(aes(x=decimal_date, y=co2_new)) + 
  geom_line()

Let’s plot CO2 of the same month as a function of decimal_date, with the color option:

Keeling_Data_tbl %>% 
  mutate(co2_new = ifelse(quality==1, co2, NA)) %>% 
  # Make the plot
  ggplot(aes(x=decimal_date, y=co2_new, color=month)) + 
  geom_line()

Or plot the same data but in panels (facets), with the facet_wrap function:

Keeling_Data_tbl %>% 
  mutate(co2_new = ifelse(quality==1, co2, NA)) %>% 
  # Make the plot
  ggplot(aes(x=decimal_date, y=co2_new, color=month)) + 
  geom_line() +
  facet_wrap(~ month)

OK, that is enough for data wrangling We will learn more about ggplot2 in future sections.

Quick plots

Now think about what will you do when get a new dataset before analyzing it?

Normally, we would like to do two things - to take a quick look at the statics and to plot the patterns.

R has many built-in functions for a large number of summary statistics. For example, mean(), median(), range(), max(), min(), sd(), var(), IQR(), and summary(). Make sure you understand how to use them and always pay attentions to NA values.

Our next task is to visualize the data. Often, a proper visualization can illuminate features of the data that can inform further analysis. We will learn three basic types of plots in this section.

Histograms

When visualizing a single numerical variable, a histogram will be our go-to tool, which can be created in R using the hist() function.

# Add a new column to the original tibble
Keeling_Data_tbl <- Keeling_Data_tbl %>% 
    mutate(co2_new = ifelse(quality==1, co2, NA))

# Notice we use pull() to get a vector from a tibble
Month_CO2 <- Keeling_Data_tbl %>% 
  pull(co2_new)

# plot hist
hist(Month_CO2)

The histogram function has a number of parameters which can be changed to make our plot look much nicer. Use the ? operator to read the documentation for the hist() to see a full list of these parameters.

hist(Month_CO2,
     xlab = "Monthly CO2 mixing ratios (ppm)",
     main = "Histogram of Monthly CO2",
     breaks = 20,
     col = "blue",
     border = "red")
box(lwd=2,col="green")

By default R will attempt to intelligently guess a good number of breaks, but as we can see here, it is sometimes useful to modify this yourself.

Boxplots

To visualize the relationship between a numerical and categorical variable, we will use a boxplot. A boxplot is usful in checking the center, spread, and skewness of a sample.

In the Keeling_Data_tbl dataset, the month is a categorical variable. First, note that we can use a single boxplot as an alternative to a histogram for visualizing a single numerical variable. To do so in R, we use the boxplot() function.

boxplot(Month_CO2)

Or try to plot the raw data:

boxplot(Keeling_Data$co2)

However, more often, we will use boxplots to compare a numerical variable for different values of a categorical variable.

boxplot(co2_new ~ month, data=Keeling_Data_tbl)

Here we used the boxplot() command to create side-by-side boxplots. However, since we are now dealing with two variables, the syntax has changed. The R syntax co2_new ~ month, data = Keeling_Data_tbl reads “Plot the co2_new variable against the month variable using the Keeling_Data_tbl dataset.” We see the use of a ~ (which specifies a formula) and also a data = argument. This will be a syntax that is common to many functions we will use later.

Again, boxplot() has a number of additional arguments which have the ability to make our plot more visually appealing.

boxplot(co2_new ~ month, data=Keeling_Data_tbl,
xlab = "Month",
ylab = "CO2 (ppm)",
main = "Monthly CO2",
cex = 2,
col = "orange",
border = "darkgreen")

Scatterplots

Lastly, to visualize the relationship between two numeric variables, we will use a scatterplot. This can be done with the plot() function and the ~ syntax we just used with a boxplot. (The function plot() can also be used more generally; see the documentation for details.)

plot(co2_new ~ year, data=Keeling_Data_tbl)

plot(co2_new ~ year, data=Keeling_Data_tbl,
xlab = "Year",
ylab = "CO2 (ppm)",
main = "CO2 vs Year",
pch = "+",
cex = 2,
col = "navy")

In-class exercises

Exercise #1

Use Keeling_Data, compute the annual mean of CO2 since 1959, plot your results.