The goal of this tutorial is to group the missing values to learn which missing values are isolated and which belongs to largue groups. This could lead to different treatment of the missing values according to different criteria.
library(dplyr)# In this tutorial we use the global power plant database
# http://datasets.wri.org/dataset/globalpowerplantdatabase
Dataset <- read.csv("global_power_plant_database.csv", header = TRUE, stringsAsFactors = FALSE)
head(Dataset)## country country_long
## 1 AFG Afghanistan
## 2 AFG Afghanistan
## 3 AFG Afghanistan
## 4 AFG Afghanistan
## 5 AFG Afghanistan
## 6 AFG Afghanistan
## name gppd_idnr
## 1 Kajaki Hydroelectric Power Plant Afghanistan GEODB0040538
## 2 Mahipar Hydroelectric Power Plant Afghanistan GEODB0040541
## 3 Naghlu Dam Hydroelectric Power Plant Afghanistan GEODB0040534
## 4 Nangarhar (Darunta) Hydroelectric Power Plant Afghanistan GEODB0040536
## 5 Northwest Kabul Power Plant Afghanistan GEODB0040540
## 6 Pul-e-Khumri Hydroelectric Power Plant Afghanistan GEODB0040537
## capacity_mw latitude longitude fuel1 fuel2 fuel3 fuel4
## 1 33.00 32.3220 65.1190 Hydro
## 2 66.00 34.5560 69.4787 Hydro
## 3 100.00 34.6410 69.7170 Hydro
## 4 11.55 34.4847 70.3633 Hydro
## 5 42.00 34.5638 69.1134 Gas
## 6 6.00 35.9416 68.7100 Hydro
## commissioning_year owner source url
## 1 NA GEODB http://globalenergyobservatory.org
## 2 NA GEODB http://globalenergyobservatory.org
## 3 NA GEODB http://globalenergyobservatory.org
## 4 NA GEODB http://globalenergyobservatory.org
## 5 NA GEODB http://globalenergyobservatory.org
## 6 NA GEODB http://globalenergyobservatory.org
## geolocation_source year_of_capacity_data generation_gwh_2013
## 1 GEODB 2017 NA
## 2 GEODB 2017 NA
## 3 GEODB 2017 NA
## 4 GEODB 2017 NA
## 5 GEODB 2017 NA
## 6 GEODB 2017 NA
## generation_gwh_2014 generation_gwh_2015 generation_gwh_2016
## 1 NA NA NA
## 2 NA NA NA
## 3 NA NA NA
## 4 NA NA NA
## 5 NA NA NA
## 6 NA NA NA
## estimated_generation_gwh
## 1 NA
## 2 NA
## 3 NA
## 4 NA
## 5 NA
## 6 NA# We are going to identify the missing values of the column generation_gwh_2016
# Percentage of missing values in this column is
round(length(which(is.na(Dataset$generation_gwh_2016)))/length(Dataset$generation_gwh_2016) * 100,2)## [1] 70.95# The index of the missing values on this column are
my_nas_index <- which(is.na(Dataset$generation_gwh_2016))# The function runs over the indexes and assign same group to consecutive indexes
group_my_nas <- function(my_na_vector){
group_number <- 1
group_vector <- c()
for(i in 1:(length(my_na_vector)-1)){
group_vector[i] <- group_number
if((my_na_vector[i+1] - my_na_vector[i]) > 1){
group_number <- group_number + 1
}
}
if(last(group_vector) != group_number){
group_vector[length(group_vector)+1] <- group_number
}
return(group_vector)
}# We apply the function to the indexes of the missing values
# We get different groups according to how the nas are grouped
head(group_my_nas(my_nas_index))## [1] 1 1 1 1 1 1# We can check the size of those groups using the table function
# We see that most of the missing values are grouped in 5 big groups of thousands
table(group_my_nas(my_nas_index))##
## 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
## 760 7652 2352 13 49 10 36 17 25 11 8 5 1 2 4
## 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
## 13 2 13 6 39 12 7 1 45 23 8 2 9 5 19
## 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45
## 1 1 1 1 2 7 9 8 2 2 8 1 6 1 7
## 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
## 23 2 23 18 15 27 1 32 2 49 1 1 44 3 39
## 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75
## 28 42 8 2 2 11 4 34 83 2 1920 1 451 15 1
## 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
## 74 2 3 4 8 3 2 6 1 1 4 1571 2 5 4198
## 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
## 1 1 1 1 2 1 1 1 1 1 1 1 1 1 2
## 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
## 1 1 1 2 1 1 1 1 1 1 2 1 1 1 1
## 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2
## 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2
## 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
## 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
## 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
## 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
## 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1
## 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
## 2 1 131 2 2 1 1 4 4 1 1 1 2 4 2
## 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255
## 3 3 1 1 5 1 1 3 1 9 1 1 1 1 2
## 256 257 258 259
## 1 1 2 23In this tutorial we have learnt how to find the distribution of missing values. The best way to deal with this is creating a function that groups the missing values if they are consecutive.