-
Notifications
You must be signed in to change notification settings - Fork 0
/
02_data_cleaning.R
198 lines (154 loc) · 5.55 KB
/
02_data_cleaning.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
library(tidyverse)
library(CoordinateCleaner)
library(lubridate)
library(assertthat)
source("R/functions.R")
data("buffland")
#==============================================================================
# Import GBIF species occurrence data
# Clean up the "gbif_occurrences" folder, as needed
files <- list.files("data/gbif_occurrences", full.names = TRUE)
cleaning.table <- tibble(file_name = files) %>%
mutate(
species = gsub("data/gbif_occurrences/", "", file_name) %>%
str_extract(., "[a-zA-Z]+_[a-z\\-]+") %>%
gsub("_", " ", .),
limit = as.numeric(str_extract(file_name, "[0-9]+"))
) %>%
arrange(species, limit)
cleaning.table2 <- cleaning.table %>%
group_by(species) %>%
count() %>%
filter(n > 1) %>%
mutate(needs_cleaning = "yes") %>%
select(-n)
cleaning.table <- left_join(cleaning.table, cleaning.table2, by = "species")
files.to.delete <- cleaning.table %>%
filter(needs_cleaning == "yes") %>%
distinct(species, .keep_all = TRUE) %>%
pull(file_name)
sapply(files.to.delete, function(x) file.remove(x))
# Pull in every species-level data frame, keeping track of which species name
# was queried using the "query_name" column
d <- data.frame()
files <- list.files("data/gbif_occurrences", full.names = TRUE)
gbif.metadata <- read_csv("data/gbif_miscellaneous/gbif_metadata.csv")
nrow(gbif.metadata) # number of species queried on GBIF
n.species.wo.data <- filter(gbif.metadata, n_records == 0) %>% nrow()
n.species.wo.data
n.species.w.data <- filter(gbif.metadata, n_records > 0) %>% nrow()
n.species.w.data
assert_that(length(files) == n.species.w.data)
d <- plyr::ldply(files, custom_read_csv)
# Arrange by the data by (GBIF) species name and rename columns
d <- d %>%
arrange(species, key) %>%
dplyr::rename(
latitude = decimalLatitude,
longitude = decimalLongitude
)
dim(d)
# Verify that the full data frame has the same number of species as we have
# occurrence files
assert_that(length(unique(d$species)) == length(files))
# Verify that the full data frame has the same number of records as the
# metadata suggests
assert_that(nrow(d) == sum(gbif.metadata$n_records))
#==============================================================================
# Data cleaning
# As a safeguard, filter out any records not from kingdom Plantae
table(d$kingdom, useNA = "ifany")
d <- filter(d, kingdom == "Plantae")
assert_that(unique(d$kingdom) == "Plantae")
# Add on IUCN assessment information
a <- read_csv("data/IUCN/assessments.csv") %>%
mutate(iucn_assessment_year = year(assessmentDate))
d <- d %>%
left_join(
.,
select(a, scientificName, redlistCategory, iucn_assessment_year),
by = c("query_name" = "scientificName")
)
# Save the unfiltered GBIF data for Data Deficient species to disk
d %>%
filter(redlistCategory == "Data Deficient") %>%
write_csv("data/gbif_miscellaneous/unfiltered_gbif_data_deficient.csv")
# Filter out unwanted "basisOfRecord" values and data points that
# are dated after the IUCN assessment was done
table(d$basisOfRecord, useNA = "ifany")
d <- d %>%
filter(
basisOfRecord %in% c("HUMAN_OBSERVATION", "OBSERVATION"),
year < iucn_assessment_year
)
table(d$basisOfRecord, useNA = "ifany")
dim(d)
# CoordinateCleaner flagging
d <- d %>%
mutate(
# Flag any invalid coordinates
val = cc_val(., lat = "latitude", lon = "longitude", value = "flagged"),
# Flag any identical coordinates
equ = cc_equ(., lat = "latitude", lon = "longitude", value = "flagged"),
# Flag any coordinates near country centroids
cen = cc_cen(., lat = "latitude", lon = "longitude",
buffer = 1000, value = "flagged"),
# Flag any coordinates near country capitals
cap = cc_cap(., lat = "latitude", lon = "longitude",
buffer = 1000, value = "flagged"),
# Flag any coordinates in the vicinity of GBIF headquarters
gbif = cc_gbif(., lat = "latitude", lon = "longitude",
buffer = 1000, value = "flagged"),
# Flag any coordinates in the vicinity of biodiversity institutions
inst = cc_inst(., lat = "latitude", lon = "longitude",
buffer = 1000, value = "flagged"),
# Flag coordinates falling over the open ocean
sea = cc_sea(., lat = "latitude", lon = "longitude",
ref = buffland, value = "flagged")
)
# CoordinateCleaner filtering
d <- d %>%
filter(
val == TRUE,
equ == TRUE,
cen == TRUE,
cap == TRUE,
gbif == TRUE,
inst == TRUE,
sea == TRUE
) %>%
select(-val, -equ, -cen, -cap, -gbif, -inst, -sea)
dim(d)
# How many plant species represented?
n_distinct(d$query_name)
n_distinct(d$species)
#==============================================================================
# Filter to only species with >= 3 occurrence points
d <- d %>%
left_join(
.,
d %>%
group_by(species) %>%
summarize(number_of_points = n()),
by = "species"
) %>%
filter(number_of_points >= 3)
dim(d)
# How many plant species represented?
n_distinct(d$query_name)
n_distinct(d$species)
# Generate and save metadata on this derived dataset for GBIF citation purposes
derived.dataset.metadata <- d %>%
group_by(datasetKey) %>%
summarize(nOccurrencesInDerivedDataset = n()) %>%
ungroup()
assertthat::assert_that(
sum(derived.dataset.metadata$nOccurrencesInDerivedDataset) == nrow(d)
)
write_csv(
derived.dataset.metadata,
"data/gbif_miscellaneous/gbif_derived_dataset_metadata.csv"
)
#==============================================================================
# Write cleaned GBIF data to disk
write_csv(d, "data/gbif_cleaned/gbif_all.csv")