-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcpod_analysis.R
261 lines (201 loc) · 15.6 KB
/
cpod_analysis.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
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
library(tidyverse)
library(lessR)
library(tidyr)
setwd("~/lifewatch_network_analysis/cpod/")
#input data
load("cpod_df_20180701_20220801_week_hourly.Rdata")
cpod1_df <- as.data.frame(cpod1_df)
season <- as.data.frame(read_csv("csv/seasons.csv"))
#recode station names
cpod1_df$station <- recode_factor(cpod1_df$station,"bpns-Reefballs Belwind"="Belwindreefballs","bpns-Reefballs-cpower"="Cpowerreefballs",
"AP_bpns-Grafton" = "Grafton", "bpns-Nautica Ena" = "Nauticaena", "bpns-Grafton"="Grafton", "bpns-G88" ="G-88",
"bpns-Faulbaums" = "Faulbaums", "bpns-Westhinder" = "Westhinder", "bpns-Buitenratel" = "Buitenratel","bpns-Gardencity" = "Gardencity",
"bpns-Birkenfels" = "Birkenfels", "AP-bpns-Birkenfels"="Birkenfels", "AP-bpns-Belwind" = "Belwindreefballs", "AP-bpns-Cpower"="Cpowerreefballs")
cpod1_df$station <- factor(as.character(cpod1_df$station), levels=rev(levels(cpod1_df$station)))
#convert time to POSIXct class
cpod1_df[,2] <- as.POSIXct(cpod1_df[,2], format = "%Y-%m-%d %H:%M:%S", tz="UTC")
season[,6] <- as.POSIXct(season[,6], format = "%Y-%m-%d %H:%M:%S", tz="UTC")
#add season
df <- cpod1_df %>% mutate(date=format(time,"%Y-%m-%d",tz="UTC"),year = as.numeric(format(time, "%Y", tz="UTC")), month = as.numeric(format(time, "%m", tz="UTC")), day = as.numeric(format(time, "%d", tz="UTC")), hour = as.numeric(format(time, "%H", tz="UTC")),
noise = number_clicks_total - number_clicks_filtered) %>%
mutate(Season = case_when (time >= season[1,6] & time < season[2,6] ~ season[1,1],
time >= season[2,6] & time < season[3,6] ~ season[2,1],
time >= season[3,6] & time < season[4,6] ~ season[3,1],
time >= season[4,6] & time < season[5,6] ~ season[4,1],
time >= season[5,6] & time < season[6,6] ~ season[5,1],
time >= season[6,6] & time < season[7,6] ~ season[6,1],
time >= season[7,6] & time < season[8,6] ~ season[7,1],
time >= season[8,6] & time < season[9,6] ~ season[8,1],
time >= season[9,6] & time < season[10,6] ~ season[9,1],
time >= season[10,6] & time < season[11,6] ~ season[10,1],
time >= season[11,6] & time < season[12,6] ~ season[11,1],
time >= season[12,6] & time < season[13,6] ~ season[12,1],
time >= season[13,6] & time < season[14,6] ~ season[13,1],
time >= season[14,6] & time < season[15,6] ~ season[14,1],
time >= season[15,6] & time < season[16,6] ~ season[15,1],
time >= season[16,6] & time < season[17,6] ~ season[16,1],
time >= season[17,6] & time < season[18,6] ~ season[17,1],
time >= season[18,6] & time < season[19,6] ~ season[18,1],
time >= season[19,6] & time < season[20,6] ~ season[19,1]))
df <- df %>% filter(station != "G-88")
###CHECKING FOR DUPLICATES
#check if there are duplicates
duplicates0 <- cpod1_df[duplicated(cpod1_df[,c(1,2,3,13,17)]),c(1,2,3,13,17)]
duplicates0[,2] <- as.POSIXct(duplicates0[,2], format = "%Y-%m-%d %H:%M:%S", tz="UTC")
duplicates <- merge(duplicates0,cpod1_df)
save(duplicates,file="csv/cpod1_df_duplicates.csv")
#remove duplicated rows where recorded = 10 mins, retain only 60 mins.
duplicates_10mins <- duplicates %>% filter(recorded==10)
df <- anti_join(df,duplicates_10mins)
#recode season
df$Season <- recode(df$Season, "September Equinox" = "Autumn","December Solstice" = "Winter", "March Equinox"="Spring", "June Solstice"="Summer")
write_csv(df, "csv/df.csv")
#double check if there are duplicates
duplicates0 <- df[duplicated(df[,c(1,2,3,13,17)]),c(1,2,3,13,17)]
duplicates0[,2] <- as.POSIXct(duplicates0[,2], format = "%Y-%m-%d %H:%M:%S", tz="UTC")
duplicates <- merge(duplicates0,df)
save(duplicates,file="csv/cpod1_df_duplicates_60mins.csv")
#---still there are duplicates both with 60 mins of recordings
###DATA EXPLORATION
df %>% ggplot(aes(time, dpm, colour = station)) + geom_point(size = 0.1)
#analysis by day, lost_hrs>9 or lost_hrs>1
df_day <- df %>% group_by(station,longitude, latitude,deployment_fk,date,time,hour,Season,species) %>% summarise(number_clicks_filtered=sum(number_clicks_filtered), number_clicks_total=sum(number_clicks_total), noise=sum(noise), dph=if_else(dpm>0,1,0), dpm=sum(dpm), lost_minutes=sum(lost_minutes)) %>%
group_by(station,longitude, latitude,deployment_fk,date,Season, species) %>% summarise(hrs = length(unique(hour)),number_clicks_filtered=sum(number_clicks_filtered), number_clicks_total=sum(number_clicks_total), noise=sum(noise), dph=sum(dph), dpm=sum(dpm), lost_minutes=sum(lost_minutes)) %>%
mutate(noise_percent = noise/number_clicks_total, dpd = if_else(dpm>0,1,0), lost_hrs = if_else(lost_minutes>9,1,0))
write_csv(df_day, "csv/df_day_lost_mins_10.csv")
#---compute station activity
cpod1_df %>% mutate(date=format(time,"%Y-%m-%d",tz="UTC")) %>% group_by(station, deployment_fk, time,date) %>% summarise(hrs_activity = length(unique(time))) %>%
group_by(station) %>% summarise(hrs_activity = sum(hrs_activity), no_deployments = length(unique(deployment_fk)), mean_hrs_deploy = hrs_activity/no_deployments, no_days=length(unique(date))) %>%
write_csv("csv/station_activity.csv")
#---visualize station activity
cpod1_df %>%
ggplot(aes(time, station)) +
geom_point(size = 0.1) +
ggtitle("Cetacean Passive Acoustic Network Station Activity") + theme_linedraw()+
theme(axis.text.y=element_text(size=10), axis.text.x = element_text(size=10), axis.title = element_blank(),
plot.title=element_text(size=8))
ggsave("plots/station_activity.png", device='png', width=5,height=5)
#NBHF clicks and noise per station
df_clicks <- cpod1_df %>% filter(species == "NBHF" | species =="Dolphins") %>% select(station, time, number_clicks_filtered,number_clicks_total) %>%
group_by(station,time) %>% summarise(number_clicks_total=mean(number_clicks_total), number_clicks_filtered=sum(number_clicks_filtered)) %>% mutate(noise = number_clicks_total - number_clicks_filtered)
df_long <- gather(df_clicks, type, number, c(number_clicks_filtered,noise), factor_key=TRUE)
df_long$type <- relevel(df_long$type,"noise")
thm <- theme_minimal() + theme(plot.title = element_text(hjust = 0.5, size =10), axis.text=element_text(size=10), legend.title=element_blank())
ggplot(df_long) + geom_line(aes(x = time, y = number,color = type), size = 0.8) + facet_wrap(~station, nrow=2) + thm + xlab("time") + ylab("number of clicks")+scale_colour_manual(values = c("steelblue", "orange"))
ggsave("plots/noise_filtered_ratio.png", device='png', width=12,height=6)
#Dolphin clicks per station
df_clicks <- df %>% filter(species =="Dolphins") %>% select(station, time, number_clicks_filtered,number_clicks_total) %>%
group_by(station,time) %>% summarise(number_clicks_total=mean(number_clicks_total), number_clicks_filtered=sum(number_clicks_filtered)) %>% mutate(noise = number_clicks_total - number_clicks_filtered)
thm <- theme_minimal() + theme(plot.title = element_text(hjust = 0.5, size =10), axis.text=element_text(size=12), legend.title=element_blank())
m <- ggplot(df_clicks) + geom_line(aes(x = time, y = number_clicks_filtered), size = 0.8) + facet_wrap(~station, nrow=2) + thm + xlab("time") + ylab("number_clicks_filtered - Dolphins")+scale_colour_manual(values = "orange")
#deployment activity
cpod1_df$deployment_fk <- as.factor(cpod1_df$deployment_fk)
cpod1_df %>%
ggplot(aes(time, deployment_fk)) +
geom_point(size = 0.1, color="orange") + theme_linedraw()+
ggtitle("Deployment Activity") + theme(axis.text.x = element_text(size=8),axis.ticks.y=element_blank(),axis.text.y=element_blank())+facet_wrap(~station,nrow=2)
ggsave("plots/deployment_activity.png", device='png', width=9,height=6)
#hours of activity per season
df %>% group_by(station,Season) %>% summarise(hrs=length(unique(time))) %>% group_by(Season) %>% summarise(hrs=sum(hrs))
#remove G-88
df_day <- df_day %>% filter(station!= "G-88")
#---boxplots of daily clicks
p_meds <- df_day%>% filter(species == "NBHF") %>% ddply(.(station), summarise, med = median(number_clicks_filtered))
df_day %>% filter(species == "NBHF") %>% ggplot(aes(x=station, y=number_clicks_filtered)) + theme_linedraw()+
geom_boxplot() +labs(y= "Daily clicks - NBHF")+ geom_text(data = p_meds, aes(x = station, y = med, label = med), size = 3, vjust = 2)
ggsave("plots/daily_clicks_nbhf.png")
p <- df_day %>% filter(species == "NBHF") %>% ggplot(aes(x=station, y=number_clicks_filtered, fill=Season)) +
geom_boxplot() +labs(y= "Daily clicks - NBHF")
ggsave("outputs/daily_clicks_nbhf_season.png")
p_meds <- df_day %>% filter(species == "Dolphins") %>% ddply(.(station), summarise, med = median(number_clicks_filtered))
p <- df_day %>% filter(species == "Dolphins") %>% ggplot(aes(x=station, y=number_clicks_filtered)) +
geom_boxplot() +labs(y= "Daily clicks - Dolphins")
ggsave("plots/daily_clicks_dol.png")
p <- df_day %>% filter(species == "Dolphins") %>% ggplot(aes(x=station, y=number_clicks_filtered, fill=Season)) +
geom_boxplot() +labs(y= "Daily clicks - Dolphins")+theme_linedraw()+theme(axis.title.x = element_blank())
ggsave("plots/daily_clicks_dol_season.png",width=10,height=6)
df_day %>% filter(species == "NBHF") %>% ggplot(aes(x=station, y=noise, fill=Season)) +
geom_boxplot() +labs(y= "Daily clicks - noise") + theme_linedraw()+theme(legend.position = "bottom", axis.title.x=element_blank())
ggsave("plots/boxplot_noise_season.png", device='png', width=9,height=6)
df_day %>% filter(species == "NBHF") %>% ggplot(aes(x=station, y=lost_minutes, fill=Season)) +
geom_boxplot() +labs(y= "Lost minutes") + theme_linedraw()+theme(legend.position = "bottom", axis.title.x=element_blank())
ggsave("plots/boxplot_lostmins_season.png", device='png', width=9,height=6)
#---boxplots of DPM
p_meds <- df_day %>% filter(species == "NBHF") %>% ddply(.(station), summarise, med = median(dpm))
df_day %>% filter(species == "NBHF") %>% ggplot(aes(x=station, y=dpm)) + theme_linedraw()+
geom_boxplot() +labs(y= "Daily Detection Positive Minutes (DPM) - NBHF")+ geom_text(data = p_meds, aes(x = station, y = med+40, label = med), size = 3, vjust = 2)
ggsave("plots/dpm_daily_nbhf.png", width = 9, height = 5)
#---------noise percent over time
ggplot(df_day, aes(x = date, y = value)) +
geom_line(aes(color = variable), size = 1) +
scale_color_manual(values = c("#00AFBB", "#E7B800")) +
theme_minimal()
#---waffle plot: Proportions
library(waffle) #remotes::install_github("hrbrmstr/waffle")
#computation of proportion of noise clicks
p <- df %>% filter(species=="NBHF") %>% group_by(station) %>%
summarise(noise_percent = sum(noise)/sum(number_clicks_total)*100, porp_percent = sum(number_clicks_filtered)/sum(number_clicks_total)*100) %>%
gather(key = "Clicks", value = "Percent", 2:3, factor_key=TRUE) %>%
mutate(Clicks = case_when(Clicks == 'noise_percent' ~ 'noise',TRUE ~ 'NBHF'))
#computation of proportion of minutes with an NBHF train per season
p<- df_day %>% filter(species=="NBHF") %>% group_by(station, Season) %>%
summarise(dpm=sum(dpm), hrs =sum(hrs)) %>%
mutate(mins_active = hrs*60, dpm_percent = round((dpm/mins_active)*100)) %>% arrange(desc(dpm_percent)) %>% mutate(mins_inactive_percent = 100-dpm_percent) %>%
gather(key = "Clicks", value = "Percent", 6:7, factor_key=TRUE) %>%
mutate(Clicks = case_when(Clicks == 'dpm_percent' ~ 'DPM',TRUE ~ 'No detection minutes'))
#computation of proportion of lost minutes over total active minutes
p <- df_day %>% filter(species=="NBHF") %>% group_by(station) %>%
summarise(lost_minutes=sum(lost_minutes), hrs =sum(hrs)) %>%
mutate(mins_active = hrs*60, lost_mins_percent = round((lost_minutes/mins_active)*100), mins_not_lost = 100-lost_mins_percent) %>%
gather(key = "Minutes", value = "Percent", 5:6, factor_key=TRUE) %>%
mutate(Minutes = case_when(Minutes == 'lost_mins_percent' ~ 'lost minutes',TRUE ~ 'total active minutes'))
##computation of proportion of minutes with a Dolphin train
p <- df_day %>% filter(species=="Dolphins") %>% group_by(station) %>%
summarise(dpm=sum(dpm), hrs =sum(hrs)) %>%
mutate(mins_active = hrs*60, dpm_percent = (dpm/mins_active)*100, mins_inactive_percent = 100-dpm_percent) %>%
gather(key = "Clicks", value = "Percent", 5:6, factor_key=TRUE) %>%
mutate(Clicks = case_when(Clicks == 'dpm_percent' ~ 'DPM - Dolphins',TRUE ~ 'No detection minutes'))
#plot! change fill according to field name of p (Clicks/Minutes)
ggplot(p, aes(fill = Minutes, values = Percent)) +
geom_waffle(n_rows = 10, size = 0.5, colour = "#ffffff", flip = TRUE) +
coord_equal() + theme_enhance_waffle() + #facet_grid(rows = vars(Season), cols = vars(station))+ theme_minimal()+
scale_fill_manual(values = c("grey", "blue"), name = "") +
facet_wrap(~reorder(station, -Percent), nrow = 2)+ theme_minimal()+
theme(axis.text=element_blank(),axis.title=element_blank(), legend.position="bottom", strip.text.x = element_text(size = 15))
ggsave("plots/lostmins_station_season_waffle.png", width = 11, height = 6, dpi=300)
#plot_grid(p1, p2,p3, p4, ncol = 2, nrow=2, labels = c("December Solstice", "March Equinox", "June Solstice", "September Equinox"), label_size = 11,label_x=0.3)
#---STATISTICAL TESTS
#normality of daily clicks
tapply(X=df_day$number_clicks_filtered, INDEX=df_day$station, FUN=shapiro.test) #--> not normal
#Kruskal-Wallis Test
kruskal.test(number_clicks_filtered ~ station, data = df_day) # significant station effect
kruskal.test(number_clicks_filtered ~ Season, data = df_day) # significant seasonal effect
kruskal.test(number_clicks_filtered ~ interaction(Season,station), data = df_day) #significant interaction effect
pairwise.wilcox.test(df_day$number_clicks_filtered, df_day$station,
p.adjust.method = "BH")
pairwise.wilcox.test(df_day$number_clicks_filtered, df_day$Season,
p.adjust.method = "BH")
pairwise.wilcox.test(df_day$number_clicks_filtered, interaction(df_day$station,df_day$Season),
p.adjust.method = "BH")
#noise
kruskal.test(noise ~ station, data = df_day) # significant station effect
pairwise.wilcox.test(df_day$noise, df_day$station,
p.adjust.method = "BH")
kruskal.test(noise ~ Season, data = df_day) # significant seasonal effect
kruskal.test(noise ~ interaction(Season,station), data = df_day) #significant interaction effect
#dpm
kruskal.test(dpm ~ station, data = df_day) # significant station effect
kruskal.test(dpm ~ Season, data = df_day) # significant seasonal effect
kruskal.test(dpm ~ interaction(Season,station), data = df_day) #significant interaction effect
#lost_mins
kruskal.test(lost_minutes ~ station, data = df_day) # significant station effect
kruskal.test(lost_minutes ~ Season, data = df_day) # significant seasonal effect
kruskal.test(lost_minutes ~ interaction(Season,station), data = df_day) #significant interaction effect
#---MAPPING DF
stn_list <- read_csv("~/lifewatch_network_analysis/fish/csv/stn_list.csv")
interp_csv <- df_day %>% filter(species=="NBHF") %>% group_by(station,Season) %>% summarise(sum(number_clicks_filtered), sum(number_clicks_total), sum(noise), sum(dph), sum(dpm), sum(lost_minutes))
#add lat long
stn_list <- read_csv("R/stn_list.csv")
interp_csv$latitude <- stn_list$latitude[match(interp_csv$station,stn_list$location_col)]
interp_csv$longitude <- stn_list$longitude[match(interp_csv$station,stn_list$location_col)]
write_csv(interp_csv, "csv/interp_nbhf_season.csv")