/
Multivariate-2017.R
349 lines (282 loc) · 12.4 KB
/
Multivariate-2017.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
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
#-------------------------------------------------------------------------------
# CLEAR WATER: Predicting Water Quality in Chicago Beaches
#
# All user-defined settings are found in this file
# Make changes below as described to manipulate the model
# The model and evaluation code is located in 30_Model.R and Functions/modelEColi.R
#
# Run this file only - all other code is pulled in by Master.R
#-------------------------------------------------------------------------------
# Load libraries and functions
source("R/00_Startup.R")
#-------------------------------------------------------------------------------
# Ingest Data
#-------------------------------------------------------------------------------
# The following .R files have been run already and are cached in Data/df.Rds
# source("R/10_LabResults.R")
# source("R/11_USGSpredictions.R")
# source("R/12_LockOpenings.R")
# source("R/13_Beach_Water_Levels.R")
# source("R/14_Weather.R")
# source("R/15_WaterQuality.R")
# source("R/20_Clean.R")
# Transform wind variables for modeling
# df$windDirectionMath <- 270 - df$windBearing
# df$windU <- df$windSpeed * cos(df$windDirectionMath)
# df$windV <- df$windSpeed * sin(df$windDirectionMath)
#
# df$windDirectionMath_hourly_1 <- 270 - df$windBearing_hourly_1
# df$windU_hourly_1 <- df$windSpeed_hourly_1 * cos(df$windDirectionMath_hourly_1)
# df$windV_hourly_1 <- df$windSpeed_hourly_1 * sin(df$windDirectionMath_hourly_1)
# df$windDirectionMath_hourly_2 <- 270 - df$windBearing_hourly_2
# df$windU_hourly_2 <- df$windSpeed_hourly_2 * cos(df$windDirectionMath_hourly_2)
# df$windV_hourly_2 <- df$windSpeed_hourly_2 * sin(df$windDirectionMath_hourly_2)
# df$windDirectionMath_hourly_3 <- 270 - df$windBearing_hourly_3
# df$windU_hourly_3 <- df$windSpeed_hourly_3 * cos(df$windDirectionMath_hourly_3)
# df$windV_hourly_3 <- df$windSpeed_hourly_3 * sin(df$windDirectionMath_hourly_3)
# df$windDirectionMath_hourly_4 <- 270 - df$windBearing_hourly_4
# df$windU_hourly_4 <- df$windSpeed_hourly_4 * cos(df$windDirectionMath_hourly_4)
# df$windV_hourly_4 <- df$windSpeed_hourly_4 * sin(df$windDirectionMath_hourly_4)
# df$windDirectionMath_hourly_5 <- 270 - df$windBearing_hourly_5
# df$windU_hourly_5 <- df$windSpeed_hourly_5 * cos(df$windDirectionMath_hourly_5)
# df$windV_hourly_5 <- df$windSpeed_hourly_5 * sin(df$windDirectionMath_hourly_5)
# df$windDirectionMath_hourly_6 <- 270 - df$windBearing_hourly_6
# df$windU_hourly_6 <- df$windSpeed_hourly_6 * cos(df$windDirectionMath_hourly_6)
# df$windV_hourly_6 <- df$windSpeed_hourly_6 * sin(df$windDirectionMath_hourly_6)
# df$windDirectionMath_hourly_7 <- 270 - df$windBearing_hourly_7
# df$windU_hourly_7 <- df$windSpeed_hourly_7 * cos(df$windDirectionMath_hourly_7)
# df$windV_hourly_7 <- df$windSpeed_hourly_7 * sin(df$windDirectionMath_hourly_7)
# df$windDirectionMath_hourly_8 <- 270 - df$windBearing_hourly_8
# df$windU_hourly_8 <- df$windSpeed_hourly_8 * cos(df$windDirectionMath_hourly_8)
# df$windV_hourly_8 <- df$windSpeed_hourly_8 * sin(df$windDirectionMath_hourly_8)
#
# df_shift_1 <- shift_previous_data(1, df)
# df_shift_2 <- shift_previous_data(2, df)
# df_shift_3 <- shift_previous_data(3, df)
#
# df <- cbind(df, df_shift_1[,584:1080])
# df <- cbind(df, df_shift_2[,584:1080])
# df <- cbind(df, df_shift_3[,584:1080])
#
# saveRDS(df, paste0(getwd(),"/Data/df-3-day.Rds"))
df <- readRDS(paste0(getwd(),"/Data/df-3-day.Rds"))
#-------------------------------------------------------------------------------
# ADD PREDICTORS
#-------------------------------------------------------------------------------
df$precipIntensity.3.day.total <- df$precipIntensity.1.daysPrior +
df$precipIntensity.2.daysPrior +
df$precipIntensity.3.daysPrior
df$precipIntensity.by.8am <- df$precipIntensity_hourly_1 +
df$precipIntensity_hourly_2 +
df$precipIntensity_hourly_3 +
df$precipIntensity_hourly_4 +
df$precipIntensity_hourly_5 +
df$precipIntensity_hourly_6 +
df$precipIntensity_hourly_7 +
df$precipIntensity_hourly_8
df$cloudCover.3.day.total <- df$cloudCover.1.daysPrior +
df$cloudCover.2.daysPrior +
df$cloudCover.3.daysPrior
df$sunlightTime <- df$sunsetTime - df$sunriseTime
df$windSpeed.3.day.total <- df$windSpeed.1.daysPrior +
df$windSpeed.2.daysPrior +
df$windSpeed.3.daysPrior
df$windSpeed.by.8am <- df$windSpeed_hourly_1 +
df$windSpeed_hourly_2 +
df$windSpeed_hourly_3 +
df$windSpeed_hourly_4 +
df$windSpeed_hourly_5 +
df$windSpeed_hourly_6 +
df$windSpeed_hourly_7 +
df$windSpeed_hourly_8
df$windU.3.day.total <- df$windU.1.daysPrior +
df$windU.2.daysPrior +
df$windU.3.daysPrior
df$windU.by.8am <- df$windU_hourly_1 +
df$windU_hourly_2 +
df$windU_hourly_3 +
df$windU_hourly_4 +
df$windU_hourly_5 +
df$windU_hourly_6 +
df$windU_hourly_7 +
df$windU_hourly_8
df$windV.3.day.total <- df$windV.1.daysPrior +
df$windV.2.daysPrior +
df$windV.3.daysPrior
df$windV.by.8am <- df$windV_hourly_1 +
df$windV_hourly_2 +
df$windV_hourly_3 +
df$windV_hourly_4 +
df$windV_hourly_5 +
df$windV_hourly_6 +
df$windV_hourly_7 +
df$windV_hourly_8
df$Water.Level.3.day.total <- df$Water.Level.1.daysPrior +
df$Water.Level.1.daysPrior +
df$Water.Level.1.daysPrior
df$DayOfWeek <- as.factor(df$DayOfWeek)
df$Obrien.Lock.Volume.3.day.total <- df$Obrien.Lock.Volume.1.daysPrior +
df$Obrien.Lock.Volume.2.daysPrior +
df$Obrien.Lock.Volume.3.daysPrior
df$CRCW.Lock.Volume.3.day.total <- df$CRCW.Lock.Volume.1.daysPrior +
df$CRCW.Lock.Volume.2.daysPrior +
df$CRCW.Lock.Volume.3.daysPrior
df$Wilmette.Lock.Volume.3.day.total <- df$Wilmette.Lock.Volume.1.daysPrior +
df$Wilmette.Lock.Volume.2.daysPrior +
df$Wilmette.Lock.Volume.3.daysPrior
#-------------------------------------------------------------------------------
# CHOOSE PREDICTORS
# Comment out the predictors that you do not want to use
#-------------------------------------------------------------------------------
# set predictors
df_model <- df[, c("Escherichia.coli", #dependent variable
"Client.ID", #beach name
## Precipitation
"precipProbability",
"precipIntensity.1.daysPrior",
"precipIntensity.3.day.total",
"precipIntensity.by.8am",
## Sunlight
"cloudCover.1.daysPrior",
"cloudCover.3.day.total",
"sunlightTime",
## Wind
"windSpeed.1.daysPrior",
"windSpeed.3.day.total",
"windSpeed.by.8am",
"windU.1.daysPrior",
"windU.3.day.total",
"windU.by.8am",
"windV.1.daysPrior",
"windV.3.day.total",
"windV.by.8am",
## Tidal levels
"moonPhase",
## Lake levels
"Water.Level",
"Water.Level.1.daysPrior",
"Water.Level.3.day.total",
## Density of humans and animals
"DayOfWeek",
"DayOfYear",
### Variables NOT cited in our paper from prior literature
## Lock openings
"Obrien.Lock.Volume.1.daysPrior",
"CRCW.Lock.Volume.1.daysPrior",
"Wilmette.Lock.Volume.1.daysPrior",
"Obrien.Lock.Volume.3.day.total",
"CRCW.Lock.Volume.3.day.total",
"Wilmette.Lock.Volume.3.day.total",
## Today's readings at selected beaches
"Calumet_DNA.Geo.Mean",
"Rainbow_DNA.Geo.Mean",
"n63rd_DNA.Geo.Mean",
"Montrose_DNA.Geo.Mean",
"South_Shore_DNA.Geo.Mean",
## Train/Test split data
"Year",
"Date",
## Removed after transformation (see next line)
"DNA.Geo.Mean"
)]
## sub DNA for E. coli. This is for validation only. This will make sure DNA days 1000 and over are 1s and under 1000 are 0s.
df_model[df_model$Year == "2017","Escherichia.coli"] <- df_model[df_model$Year == "2017","DNA.Geo.Mean"] - 765
df_model$DNA.Geo.Mean <- NULL
finaltest <- df_model[df_model$Year == "2017",]
#-------------------------------------------------------------------------------
# CHOOSE TEST/TRAIN SETS
# You can decide whether to use kFolds cross validation or define your own sets
# If you set kFolds to TRUE, the data will be separated into 10 sets
# If you set kFolds to FALSE, the model will use trainStart, trainEnd, etc. (see below)
# CANNOT BE USED IF productionMode = TRUE
#-------------------------------------------------------------------------------
kFolds <- FALSE #If TRUE next 2 lines will not be used but cannot be commented out
testYears <- c("2017")
trainYears <- c("2006", "2007", "2008", "2009","2010", "2011", "2012", "2013", "2014", "2015", "2016")
# trainYears <- trainYears[! trainYears %in% testYears]
# If productionMode is set to TRUE, a file named model.Rds will be generated
# Its used is explained at https://github.com/Chicago/clear-water-app
# Set trainYears to what you would like the model to train on
# testYears must still be specified, although not applicable
# plots will not be accurate
productionMode <- FALSE
#-------------------------------------------------------------------------------
# DOWNSAMPLING
# If you set downsample to TRUE, choose the 3 variables below
# The training set will be a 50/50 split of 1) data less than the "lowMax" and
# 2) data between the "highMin" and "highMax"
#-------------------------------------------------------------------------------
# downsample settings
downsample <- FALSE #If FALSE comment out the next 3 lines
highMin <- 235
highMax <- 2500
lowMax <- 235
#-------------------------------------------------------------------------------
# EXCLUDE ENTIRE BEACHES FROM THE TEST SET
# This is important if you use same-day beach test results as a predictor
# If so, the predictor beach should not be a beach that is being predicted
# because the model would then be predicting on data it was trained on.
# Comment out any beach that you used as a predictor.
#-------------------------------------------------------------------------------
excludeBeaches <- c(
# "12th",
# "31st",
# "39th",
# "57th",
"63rd",
# "Albion",
"Calumet",
# "Foster",
# "Howard",
# "Jarvis",
"Juneway",
# "Leone",
"Montrose",
# "North Avenue",
# "Oak Street",
# "Ohio",
# "Osterman",
"Rainbow",
# "Rogers",
"South Shore"
)
#-------------------------------------------------------------------------------
# NAME PLOTS
# These are automatically generated based on the settings chosen above
#-------------------------------------------------------------------------------
title1 <- paste0("ROC",
if(kFolds == TRUE) " - kFolds",
if(kFolds == FALSE) " - validate on ",
if(kFolds == FALSE) testYears)
title2 <- paste0("PR Curve",
if(kFolds == TRUE) " - kFolds",
if(kFolds == FALSE) " - validate on ",
if(kFolds == FALSE) testYears)
#-------------------------------------------------------------------------------
# THRESHHOLD
# These settings can be used to manipulate the plots and the model_summary dataframe
#-------------------------------------------------------------------------------
threshBegin <- 1
threshEnd <- 1000
thresh <- 235
#-------------------------------------------------------------------------------
# RUN MODEL
# Plots will generate and results will be saved in "model_summary"
#-------------------------------------------------------------------------------
# runs all modeling code
source("R/30_Model.R", print.eval=TRUE)
# creates a data frame with all model results
# this aggregates the folds to generate one single curve
# for user-defined test set, this doesn't have any effect
model_summary <- plot_data %>%
group_by(thresholds) %>%
summarize(tpr = mean(tpr),
fpr = mean(fpr),
precision = mean(precision, na.rm = TRUE),
recall = mean(recall),
tp = mean(tp),
fn = mean(fn),
tn = mean(tn),
fp = mean(fp)
)
saveRDS(model, paste0("models/", "Multivariate-2017", ".Rds"))