a. 2: Demo
b. 2: Slides
You can access the full slideshow used in the 2-Tidyverse narration here.
The dataset called ‘fishtibble.csv’ can be downloaded here.
The dataset called ‘coralreefherbivores.csv’ can be downloaded here.
The dataset called ‘fish_abundance.csv’ can be downloaded here.
c. 2: Exercises
Read in the coralreefherbivores.csv dataset (if you haven’t already done so).
The “coralreefherbivores.csv” dataset contains trait information on 93 species of herbivorous coral reef fishes from the Great Barrier Reef, Australia. The first four columns provide taxonomic information (family, genus, species, and a combined genus + species column called “genspe”). The next column “sl” indicates the standard length of the fish. Then, you have three columns with morphological measurements of body depth, snout length, and eye diameter. Finally, you have two columns that indicate maximum body size (categories from XS to XL) and their schooling behavior.
Part I
- Using the herbivore dataset, perform the following actions:
- Determine the mean standard length and standard deviation for the families Acanthuridae, Labridae, and Siganidae.
- Create a variable that combines size and schooling into one column.
- Examine the mean eye diameter of the genera Zebrasoma and Scarus.
- Count the number of species in each size class.
- Calculate the ratio of snout length to standard length for each species, then determine which genus has the longest mean ratio of snout length to standard length.
- learn about and apply the following functions:
- unite()
- separate()
- distinct()
- drop_na()
Part II
The “fish_abundance.csv” dataset contains reef fish abundance information, which was collected by the Reef Life Survey around Lizard Island, Australia. It includes information on the respective survey ID and its metadata (e.g. site, latitude, longitude, date, and depth). It also includes a row for each observed species on a given visual survey, including taxonomic information (family, genus, species, and “genspe”), as well as the number of individuals observed on that survey (‘total’).
- Using the fish abundance dataset, perform the following actions:
- Reveal how many distinct surveys were performed aruond Lizard Island.
- In the family column, replace all occurrences of Scaridae with Labridae.
- Filter the dataset to only retain the families Acanthuridae, Siganidae, Kyphosidae and Labridae.
- Examine how many different species are in each genus.
- Calculate the mean abundance of each genus across different sites using group_by() and summarize(), while also retaining a variable that indicates the number of species per genus.
- Using the herbivore trait dataset, perform the following actions:
- Determine whether the median eye diameter differs in different social groups.
- Remove the standard length, size, and schooling categories dataset.
- Calculate the mean body depth, snout length, and eye diameter of each genus.
- Turn each trait into a separate column.
- Combine the two datasets:
- Join the coralreefherbivores and fish_abundance datasets in their modified versions, retaining only genera that are present in the coralreefherbivores dataset and using “genus” as the joining variable.
- Create a new variable that separates genera into high (>=2.0) and low (<2.0) abundance genera
- Examine whether high and low abundance genera differ in their mean morphological traits
d. 2: Solutions
Part I
- Using the herbivore dataset, perform the following actions:
- Determine the mean standard length and standard deviation for the families Acanthuridae, Labridae, and Siganidae.
- Create a variable that combines size and schooling into one column.
- Examine the mean eye diameter of the genera Zebrasoma and Scarus.
- Count the number of species in each size class.
- Calculate the ratio of snout length to standard length for each species, then determine which genus has the longest mean ratio of snout length to standard length.
herbivores <- read.csv(file = "data/coralreefherbivores.csv")
herbivores.filter <- herbivores %>% # mean sl and sd sl
filter(family == c("Acanthuridae", "Labridae", "Siganidae")) %>%
group_by(family) %>%
summarize(mean.sl = mean(sl), sd.sl = sd(sl))
head(herbivores.filter)# A tibble: 3 × 3
family mean.sl sd.sl
<chr> <dbl> <dbl>
1 Acanthuridae 186. 89.7
2 Labridae 269. 73.5
3 Siganidae 168. 54.7herbivores.size_schooling <- herbivores %>%
mutate(size_school = paste(size, schooling, sep = "."))
head(herbivores.size_schooling) family genus species gen.spe
1 Acanthuridae Acanthurus achilles Acanthurus.achilles
2 Acanthuridae Acanthurus albipectoralis Acanthurus.albipectoralis
3 Acanthuridae Acanthurus auranticavus Acanthurus.auranticavus
4 Acanthuridae Acanthurus blochii Acanthurus.blochii
5 Acanthuridae Acanthurus dussumieri Acanthurus.dussumieri
6 Acanthuridae Acanthurus fowleri Acanthurus.fowleri
sl bodydepth snoutlength eyediameter size schooling
1 163.6667 0.5543625 0.4877797 0.3507191 S Solitary
2 212.7300 0.4405350 0.4402623 0.2560593 M SmallGroups
3 216.0000 0.4726556 0.5386490 0.2451253 M MediumGroups
4 82.9000 0.5586486 0.4782217 0.3196155 M SmallGroups
5 193.7033 0.5457248 0.5661867 0.2807218 L Solitary
6 266.0000 0.4669521 0.5950563 0.2217376 M Solitary
size_school
1 S.Solitary
2 M.SmallGroups
3 M.MediumGroups
4 M.SmallGroups
5 L.Solitary
6 M.Solitaryherbivores.size_schooling2 <- unite(herbivores, col = "size_and_schooling", c("size", "schooling"), sep = ".")
head(herbivores.size_schooling2) family genus species gen.spe
1 Acanthuridae Acanthurus achilles Acanthurus.achilles
2 Acanthuridae Acanthurus albipectoralis Acanthurus.albipectoralis
3 Acanthuridae Acanthurus auranticavus Acanthurus.auranticavus
4 Acanthuridae Acanthurus blochii Acanthurus.blochii
5 Acanthuridae Acanthurus dussumieri Acanthurus.dussumieri
6 Acanthuridae Acanthurus fowleri Acanthurus.fowleri
sl bodydepth snoutlength eyediameter size_and_schooling
1 163.6667 0.5543625 0.4877797 0.3507191 S.Solitary
2 212.7300 0.4405350 0.4402623 0.2560593 M.SmallGroups
3 216.0000 0.4726556 0.5386490 0.2451253 M.MediumGroups
4 82.9000 0.5586486 0.4782217 0.3196155 M.SmallGroups
5 193.7033 0.5457248 0.5661867 0.2807218 L.Solitary
6 266.0000 0.4669521 0.5950563 0.2217376 M.Solitary#alternative solution
zebrasoma.scarus <- herbivores %>%
filter(genus == c("Zebrasoma", "Scarus")) %>%
group_by(genus) %>%
summarize(mean.eye = mean(eyediameter))
zebrasoma.scarus <- herbivores %>%
filter(genus %in% c("Zebrasoma", "Scarus")) %>%
group_by(genus) %>%
summarize(mean.eye = mean(eyediameter))
number.species <- herbivores %>% # solution 1
select(size) %>%
group_by(size) %>%
count(size)
number.species# A tibble: 5 × 2
# Groups: size [5]
size n
<chr> <int>
1 L 18
2 M 43
3 S 28
4 XL 5
5 XS 2crh.d <- herbivores %>%
group_by(size) %>%
summarize(n_species = n_distinct(gen.spe))
crh.d# A tibble: 5 × 2
size n_species
<chr> <int>
1 L 18
2 M 43
3 S 28
4 XL 5
5 XS 2ratio <- herbivores %>%
group_by(species) %>%
mutate(ratio.snout.sl = (snoutlength/sl)) %>%
ungroup() %>%
group_by(genus) %>%
summarize(mean.ratio = mean(ratio.snout.sl)) %>%
arrange(desc(mean.ratio))
head(ratio)# A tibble: 6 × 2
genus mean.ratio
<chr> <dbl>
1 Zebrasoma 0.00519
2 Paracanthurus 0.00509
3 Ctenochaetus 0.00492
4 Acanthurus 0.00320
5 Siganus 0.00239
6 Naso 0.00216Part II
- Using the fish abundance dataset, perform the following actions:
- Reveal how many distinct surveys were performed around Lizard Island (SurveyID).
- In the family column, replace all occurrences of Scaridae with Labridae.
- Filter the dataset to only retain the families Acanthuridae, Siganidae, Kyphosidae and Labridae.
- Examine how many different species are in each genus.
- Calculate the mean abundance of each genus using group_by() and summarize(), while also retaining a variable that indicates the number of species per genus.
fish.abu <- read.csv(file = "data/fish_abundance.csv")
# how many surveys
surveys <- fish.abu %>%
select(surveyid) %>%
distinct() # distinct gives you the number of unique occurrences
nrow(surveys)[1] 62# replace Scaridae with Labridae
fish.abu.recode <- fish.abu %>%
mutate(family = recode(family, "Scaridae" = "Labridae")) # use recode within mutate to change the family
head(fish.abu.recode) surveyid country site sitelat sitelong surveydate
1 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
2 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
3 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
4 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
5 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
6 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
depth family genus species block total
1 3 Acanthuridae Acanthurus grammoptilus 2 4
2 3 Acanthuridae Acanthurus nigricauda 2 3
3 3 Acanthuridae Acanthurus olivaceus 2 2
4 3 Acanthuridae Ctenochaetus binotatus 1 15
5 3 Acanthuridae Ctenochaetus binotatus 2 10
6 3 Acanthuridae Zebrasoma scopas 1 2
genspe
1 Acanthurus.grammoptilus
2 Acanthurus.nigricauda
3 Acanthurus.olivaceus
4 Ctenochaetus.binotatus
5 Ctenochaetus.binotatus
6 Zebrasoma.scopas# filter out families
fish.abu.filtered <- fish.abu.recode %>%
filter(family %in% c("Acanthuridae", "Siganidae", "Labridae", "Kyphosidae")) # use filter - also works as family == c("")
head(fish.abu.filtered) surveyid country site sitelat sitelong surveydate
1 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
2 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
3 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
4 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
5 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
6 4000720 Australia Watsons Bay north -14.66 145.45 6/27/10 14:00
depth family genus species block total
1 3 Acanthuridae Acanthurus grammoptilus 2 4
2 3 Acanthuridae Acanthurus nigricauda 2 3
3 3 Acanthuridae Acanthurus olivaceus 2 2
4 3 Acanthuridae Ctenochaetus binotatus 1 15
5 3 Acanthuridae Ctenochaetus binotatus 2 10
6 3 Acanthuridae Zebrasoma scopas 1 2
genspe
1 Acanthurus.grammoptilus
2 Acanthurus.nigricauda
3 Acanthurus.olivaceus
4 Ctenochaetus.binotatus
5 Ctenochaetus.binotatus
6 Zebrasoma.scopas# count species in each genus
fish.species.counts <- fish.abu.filtered %>%
group_by(genus) %>% # group by genus
summarize(number.species = n_distinct(species)) # use n_distinct() to count the rows in each groups ;-)
head(fish.species.counts)# A tibble: 6 × 2
genus number.species
<chr> <int>
1 Acanthurid 1
2 Acanthurus 9
3 Anampses 2
4 Bodianus 3
5 Calotomus 1
6 Cetoscarus 1# mean abundances and number of species
fish.genus.abundance <- fish.abu.filtered %>%
group_by(genus) %>% # group by family and genus - important because we'll want family in this dataset
summarize(mean.abu = mean(total), number.species = n_distinct(species)) # mean and n_distinct
head(fish.genus.abundance)# A tibble: 6 × 3
genus mean.abu number.species
<chr> <dbl> <int>
1 Acanthurid 1 1
2 Acanthurus 4.01 9
3 Anampses 3.33 2
4 Bodianus 1.32 3
5 Calotomus 1 1
6 Cetoscarus 1 1- Using the herbivore trait dataset, perform the following actions:
- Determine whether the median eye diameter differs among different social groups.
- Remove the standard length, size, and schooling categories from the dataset.
- Calculate the mean body depth, snout length, and eye diameter of each genus.
- Turn each trait back into a separate column.
# eye diameter
herbivore.eyes <- herbivores %>%
ungroup() %>%
group_by(schooling) %>% # group by schooling variable
summarize(median.eye = median(eyediameter)) #summarize()
herbivore.eyes# A tibble: 5 × 2
schooling median.eye
<chr> <dbl>
1 LargeGroups 0.275
2 MediumGroups 0.230
3 Pairs 0.303
4 SmallGroups 0.261
5 Solitary 0.272herbivore.pruned <- herbivores %>%
select(-sl, -size, -schooling) # use select to remove columns
head(herbivore.pruned) family genus species gen.spe
1 Acanthuridae Acanthurus achilles Acanthurus.achilles
2 Acanthuridae Acanthurus albipectoralis Acanthurus.albipectoralis
3 Acanthuridae Acanthurus auranticavus Acanthurus.auranticavus
4 Acanthuridae Acanthurus blochii Acanthurus.blochii
5 Acanthuridae Acanthurus dussumieri Acanthurus.dussumieri
6 Acanthuridae Acanthurus fowleri Acanthurus.fowleri
bodydepth snoutlength eyediameter
1 0.5543625 0.4877797 0.3507191
2 0.4405350 0.4402623 0.2560593
3 0.4726556 0.5386490 0.2451253
4 0.5586486 0.4782217 0.3196155
5 0.5457248 0.5661867 0.2807218
6 0.4669521 0.5950563 0.2217376# mean body depth, snout length, and eye diameter
herbivore.means <- herbivore.pruned %>%
gather(5:7, key = "metric", value = "value") %>% # gather thre three morphometric columns into one
group_by(genus, metric) %>% # group by
summarize(mean.val = mean(value)) # get mean
head(herbivore.means)# A tibble: 6 × 3
# Groups: genus [2]
genus metric mean.val
<chr> <chr> <dbl>
1 Acanthurus bodydepth 0.508
2 Acanthurus eyediameter 0.298
3 Acanthurus snoutlength 0.498
4 Bolbometopon bodydepth 0.437
5 Bolbometopon eyediameter 0.233
6 Bolbometopon snoutlength 0.312# turn traits back into separate columns
herbivore.spread <- herbivore.means %>%
spread(key = "metric", value = "mean.val") # use spread function
herbivore.spread# A tibble: 14 × 4
# Groups: genus [14]
genus bodydepth eyediameter snoutlength
<chr> <dbl> <dbl> <dbl>
1 Acanthurus 0.508 0.298 0.498
2 Bolbometopon 0.437 0.233 0.312
3 Calotomus 0.385 0.246 0.306
4 Cetoscarus 0.387 0.139 0.428
5 Chlorurus 0.400 0.187 0.366
6 Ctenochaetus 0.532 0.295 0.535
7 Hipposcarus 0.404 0.150 0.386
8 Kyphosus 0.479 0.181 0.164
9 Leptoscarus 0.334 0.209 0.297
10 Naso 0.386 0.274 0.547
11 Paracanthurus 0.490 0.295 0.533
12 Scarus 0.391 0.195 0.341
13 Siganus 0.443 0.329 0.352
14 Zebrasoma 0.575 0.305 0.657- Combine the two datasets:
- Join the coralreefherbivores and fish_abundance datasets in their modified versions, retaining only genera that are present in the coralreefherbivores dataset and using “genus” as the joining variable.
- Create a new variable that separates genera into high (>=2.0) and low (<2.0) abundance genera
- Examine whether high and low abundance genera differ in their mean morphological traits
herbivores.joined <- herbivore.spread %>%
left_join(fish.genus.abundance, by = "genus") %>% # use left_join to retain the ones in the traits dataset
drop_na(mean.abu) # remove NA values
herbivores.joined# A tibble: 11 × 6
# Groups: genus [11]
genus bodydepth eyediameter snoutlength mean.abu number.species
<chr> <dbl> <dbl> <dbl> <dbl> <int>
1 Acanthur… 0.508 0.298 0.498 4.01 9
2 Calotomus 0.385 0.246 0.306 1 1
3 Cetoscar… 0.387 0.139 0.428 1 1
4 Chlorurus 0.400 0.187 0.366 5.39 2
5 Ctenocha… 0.532 0.295 0.535 5.58 3
6 Kyphosus 0.479 0.181 0.164 1.5 1
7 Naso 0.386 0.274 0.547 1.54 6
8 Paracant… 0.490 0.295 0.533 2 1
9 Scarus 0.391 0.195 0.341 3.42 15
10 Siganus 0.443 0.329 0.352 2.21 11
11 Zebrasoma 0.575 0.305 0.657 2.80 2herbivores.joined.2 <- herbivores.joined %>%
mutate(highvslow = case_when(mean.abu >=2 ~ "high", # use case_when function
TRUE ~ "low"))
herbivores.joined.2# A tibble: 11 × 7
# Groups: genus [11]
genus bodydepth eyediameter snoutlength mean.abu number.species
<chr> <dbl> <dbl> <dbl> <dbl> <int>
1 Acanthur… 0.508 0.298 0.498 4.01 9
2 Calotomus 0.385 0.246 0.306 1 1
3 Cetoscar… 0.387 0.139 0.428 1 1
4 Chlorurus 0.400 0.187 0.366 5.39 2
5 Ctenocha… 0.532 0.295 0.535 5.58 3
6 Kyphosus 0.479 0.181 0.164 1.5 1
7 Naso 0.386 0.274 0.547 1.54 6
8 Paracant… 0.490 0.295 0.533 2 1
9 Scarus 0.391 0.195 0.341 3.42 15
10 Siganus 0.443 0.329 0.352 2.21 11
11 Zebrasoma 0.575 0.305 0.657 2.80 2
# ℹ 1 more variable: highvslow <chr>herbivores.highlow <- herbivores.joined.2 %>%
gather(2:5, key = "metric", value = "value") %>% # gathering to make it more efficient
group_by(metric, highvslow) %>% # grouping
summarize(mean.val = mean(value), # summarizing for mean and sd
sd.val = sd(value))
herbivores.highlow# A tibble: 8 × 4
# Groups: metric [4]
metric highvslow mean.val sd.val
<chr> <chr> <dbl> <dbl>
1 bodydepth high 0.477 0.0688
2 bodydepth low 0.409 0.0466
3 eyediameter high 0.272 0.0567
4 eyediameter low 0.210 0.0613
5 mean.abu high 3.63 1.44
6 mean.abu low 1.26 0.299
7 snoutlength high 0.469 0.119
8 snoutlength low 0.361 0.164