-
Notifications
You must be signed in to change notification settings - Fork 0
/
04_a_preprocessing.R
210 lines (133 loc) · 6.31 KB
/
04_a_preprocessing.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
# New DATASET application
# rm(list = ls())
source('00_system.R')
ifelse(length(which(installed.packages()[,1] == 'ncdf4')) == 0 ,
install.packages('ncdf4'), print('ncdf4 installed'))
path <- paste0('Datasets/')
file_prec <- 'precipitation_monthly-mean_era5-to-1km_1979-2018-mean_v1.0.nc'
file_wind <- 'wind-speed_monthly-mean_era5-to-1km_1979-2018-mean_v1.0.nc'
file_cc <- 'cloud-cover_monthly-mean_era5-to-1km_1979-2018-mean_v1.0.nc'
fil_BIO_04 <- 'BIO04_era5-to-1km_1979-2018-mean_v1.0.nc'
file_BIO_15 <- 'BIO15_era5-to-1km_1979-2018-mean_v1.0.nc'
file_wind_meridional <- 'meridional-wind-speed_monthly-mean_era5-to-1km_1979-2018-median_v1.0.nc'
nc_prec <- nc_open(paste0(path, file_prec), auto_GMT = T)
lati <- ncvar_get(nc_prec, 'latitude')
long <- ncvar_get(nc_prec, 'longitude')
prec <- ncvar_get(nc_prec, 'precipitation_monthly-mean')
full_df <- cbind.data.frame(expand.grid('long' = long, 'lati' = lati),
'prec' = c(prec[, , 1]))
rm(prec, long, lati)
nc_wind <- nc_open(paste0(path, file_wind), auto_GMT = T)
nc_cc <- nc_open(paste0(path, file_cc), auto_GMT = T)
nc_BIO_04 <- nc_open(paste0(path, fil_BIO_04), auto_GMT = T)
nc_BIO_15 <- nc_open(paste0(path, file_BIO_15), auto_GMT = T)
nc_wind_meri <- nc_open(paste0(path, file_wind_meridional), auto_GMT = T)
wind <- ncvar_get(nc_wind, 'wind-speed')
full_df <- cbind.data.frame(full_df,
'wind' = c(wind[, , 1]))
rm(wind)
cloud_cover <- ncvar_get(nc_cc, 'cloud-cover')
full_df <- cbind.data.frame(full_df,
'cloud_c' = c(cloud_cover[,,1]))
rm(cloud_cover)
BIO_04 <- ncvar_get(nc_BIO_04, 'BIO04')
full_df <- cbind.data.frame(full_df,
'BIO04' = c(BIO_04)
)
rm(BIO_04)
BIO_15 <- ncvar_get(nc_BIO_15, 'BIO15')
full_df <- cbind.data.frame(full_df,
'BIO15' = c(BIO_15)
)
rm(BIO_15)
wind_mer <- ncvar_get(nc_wind_meri, 'meridional-wind-speed')
full_df <- cbind.data.frame(full_df,
'new_wind' = c(wind_mer[,,1])
)
rm(wind_mer)
full_df$index <- 1:dim(full_df)[1]
# first filtering
index_to_swiss <- which(full_df$long < 10.75 & full_df$long > 5 &
full_df$lati < 48 & full_df$lati > 45.75)
swiss_df <- full_df[index_to_swiss, ]
poly_sf <- st_as_sf(swiss_df, coords = c('long','lati'), crs = st_crs("+proj=longlat +datum=WGS84 +no_defs"))
country_boundary <- st_as_sf(geodata::gadm("Switzerland", level=1, path = "."))
#
country_boundary_transf <- st_transform(country_boundary, crs = st_crs("+proj=longlat +datum=WGS84 +no_defs"))
intersection_step <- st_intersection(poly_sf, country_boundary_transf)
df_step <- cbind.data.frame(as.data.frame(intersection_step),
st_coordinates(intersection_step))
df_step <- df_step[, -c(8:19)]
colnames(df_step)[8:9] <- c("long", "lati")
# TO get the elevation
swiss_elev <- df_step
coordinates(swiss_elev) <- c('long', 'lati')
proj4string(swiss_elev) <- ("+proj=longlat +datum=WGS84 +no_defs")
eee <- st_as_sf(swiss_elev)
aaa <- elevatr::get_elev_point(eee, units = 'meters',
src = 'aws', z = 7)
swiss_finalll <-as.data.frame(aaa)
df_step$elevation <- swiss_finalll$elevation
df_step <- df_step[,c(8,9,1,2,3,4,5,6,10)]
df_step$log_elevation <- log(df_step$elevation)
df_step$log_wind <- log(df_step$wind)
df_step$log_cloud <- log(df_step$cloud_c)
df_step$log_lati <- log(df_step$lati)
df_step$log_long <- log(df_step$long)
df_step$prec <- df_step$prec * 60 * 60 * 24 * 1000
# Create subsets
set.seed(181222)
# Create clusters
if(T){
# Cluster based on longitude, latitude, and cloud coverage
to_test <-as.data.frame(scale(df_step[,c(1,2,5)]))
a_ver_che <- kmeans(to_test,centers = 475,iter.max = 50)
to_check <- sample(500,50)
striped_data_b <- which(a_ver_che$cluster %in% to_check)
}
# Create stripes
if(T){
angle_deg <- 10
# Function to rotate coordinates
rotate_coords <- function(lat, long, angle_deg) {
angle_rad <- angle_deg * pi / 180
lat_rot <- lat * cos(angle_rad) - long * sin(angle_rad)
long_rot <- lat * sin(angle_rad) + long * cos(angle_rad)
return(list(lat_rot = lat_rot, long_rot = long_rot))
}
# Rotate the DataFrame coordinates
coords_rotated <- rotate_coords(df_step$lati, df_step$long, angle_deg)
df_step$lat_rot <- coords_rotated$lat_rot
df_step$long_rot <- coords_rotated$long_rot
strip_width <- 0.005 # width of each stripe and gap
strips <- cut(df_step$long_rot, breaks=seq(min(df_step$long_rot), max(df_step$long_rot), by=strip_width))
strip_width <- 0.01 # width of each stripe and gap
strips_2 <- cut(df_step$long_rot, breaks=seq(min(df_step$long_rot), max(df_step$long_rot), by=strip_width))
strip_width <- 0.05 # width of each stripe and gap
strips_3 <- cut(df_step$long_rot, breaks=seq(min(df_step$long_rot), max(df_step$long_rot), by=strip_width))
strip_width <- 0.1 # width of each stripe and gap
strips_4 <- cut(df_step$long_rot, breaks=seq(min(df_step$long_rot), max(df_step$long_rot), by=strip_width))
striped_data <- which(strips %in% levels(strips)[sample(length(levels(strips)),40)] |
strips_2 %in% levels(strips_2)[sample(length(levels(strips_2)),20)] |
strips_3 %in% levels(strips_3)[sample(length(levels(strips_3)),2)] |
strips_4 %in% levels(strips_4)[sample(length(levels(strips_4)),2)])
striped_data_f <- unique(c(striped_data, striped_data_b))
print(length(striped_data_f))
aaa <- sample(69965, 20000 - length(striped_data_f))
striped_data_ff <- unique(c(striped_data_f, aaa))
all_dfs <- list()
df_step <- df_step[, -c(15, 16)]
all_dfs[[1]] <- df_step[striped_data_ff, ]
# training_sample
df_remaining_two <- df_step[-striped_data_ff, ]
# random_sample <- sample(dim(df_remaining_two)[1], size = 500)
seq_sample <- floor(seq(1, dim(df_remaining_two)[1], length.out = 500))
all_dfs[[2]] <- df_remaining_two[seq_sample, ]
df_reamining_three <- df_remaining_two[-seq_sample, ]
# for tapering
seq_sample_sparse <- floor(seq(1,dim(df_reamining_three)[1], length.out = 10000))
all_dfs[[3]] <- df_reamining_three[seq_sample_sparse, ]
mean(all_dfs[[1]]$prec) - mean(all_dfs[[2]]$prec)
all_dfs[[4]] <- df_step
save(all_dfs, file = 'Datasets/app_dataset.RData')
}