-
Notifications
You must be signed in to change notification settings - Fork 0
/
helper.py
324 lines (271 loc) · 15.1 KB
/
helper.py
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
#import libraries
import pandas as pd
import numpy as np
import requests
from io import StringIO
import ssl
import json
from pprint import pprint
import requests
import os
from pandas import json_normalize
from sklearn.experimental import enable_iterative_imputer
from sklearn.impute import IterativeImputer
from datetime import datetime
from sklearn.feature_selection import RFE
from sklearn.tree import DecisionTreeRegressor
ssl._create_default_https_context = ssl._create_unverified_context
def load_csv(url):
#orig_url=url
file_id = url#'https://drive.google.com/uc?export=download&id='+orig_url.split('/')[-2]
dfs = pd.read_csv(file_id,error_bad_lines=False)
return(dfs)
def load_data(cityname="Montreal"):
city_url="https://enterprisemcgill.s3.amazonaws.com/city_attributes.csv"
city=load_csv(city_url)
weather_desc_url="https://enterprisemcgill.s3.amazonaws.com/weather_description.csv"
weather_desc=load_csv(weather_desc_url)
humidity_url="https://enterprisemcgill.s3.amazonaws.com/humidity.csv"
humidity=load_csv(humidity_url)
pressure_url="https://enterprisemcgill.s3.amazonaws.com/pressure.csv"
pressure=load_csv(pressure_url)
temp_url="https://enterprisemcgill.s3.amazonaws.com/temp.csv"
temp=load_csv(temp_url)
wind_direction_url="https://enterprisemcgill.s3.amazonaws.com/wind_direction.csv"
wind_direction=load_csv(wind_direction_url)
wind_speed_url="https://enterprisemcgill.s3.amazonaws.com/wind_speed.csv"
wind_speed=load_csv(wind_speed_url)
city.columns="city"+city.columns
weather_desc.columns="weather"+weather_desc.columns
humidity.columns="humidity"+humidity.columns
pressure.columns="pressure"+pressure.columns
temp.columns="temp"+temp.columns
wind_direction.columns="wind_direction"+wind_direction.columns
wind_speed.columns="wind_speed"+wind_speed.columns
data=pd.concat([city,weather_desc,humidity,pressure,temp,wind_direction,wind_speed],axis=1)
allcitydata=data[data.columns[data.columns.str.contains(cityname+"|weatherdate")]]
allcitydata.columns= ['datetime','Description','Humidity','Wind Direction','Temperature','Pressure','Wind Speed']
#allcitydata = allcitydata.drop("city",axis=1)
return(city,weather_desc,humidity,pressure,temp,wind_direction,wind_speed,allcitydata)
def api(city):
API_KEY=os.getenv('WEATHER_API')
r = requests.get('http://api.openweathermap.org/data/2.5/weather?q="+city+"&APPID=484925ef3995e2026aa4d24818ac18b1')
response=json.loads(r.text)
response_cols=response.keys()
data=json_normalize(json.loads(r.text))
weather=json_normalize(json.loads(r.text),record_path=['weather'])
weather.columns='weather.'+weather.columns
final_data=pd.concat([data[data.columns[1:len(data.columns)]],weather],axis=1)
return(final_data)
def load_and_preprocess(cityname="Montreal"):
city,weather_desc,humidity,pressure,temp,wind_direction,wind_speed,montreal = load_data(cityname=cityname)
montreal=impute(montreal)
montreal=clean_description(montreal)
montreal=seasons(montreal)
return(montreal)
##########NEW##########
#takes the montreal df as input and imputes the columns
def impute(df):
df = df.drop(0) #dropping first row of nulls
for col in list(df.drop(['datetime', 'Description'], axis = 1).columns): #run iterative imputer on each column
imp = IterativeImputer(random_state = 6)
df[[col]] = imp.fit_transform(df[[col]])
return df
#NEW - preprocessing for the description weather column - from Steven
def clean_description(df):
#too much detail, simplifying a bit and cleaning/standarizing word notation
df['Description'] = df['Description'].str.replace(' with ', ' ')
df['Description'] = df['Description'].str.replace(' and ', ' ')
df['Description'] = df['Description'].str.replace('proximity ', '')
df['Description'] = df['Description'].str.replace('light intensity', 'light')
df['Description'] = df['Description'].str.replace('heavy intensity', 'heavy')
df['Description'] = df['Description'].str.replace('very heavy', 'heavy')
#exceptions / categories that the same
df['Description'] = df['Description'].str.replace('light drizzle rain', 'light rain')
df['Description'] = df['Description'].str.replace('light drizzle', 'light rain')
df['Description'] = df['Description'].str.replace('drizzle', 'light rain')
df['Description'] = df['Description'].str.replace('sleet', 'snow')
df['Description'] = df['Description'].str.replace('freezing', 'snow')
df['Description'] = df['Description'].str.replace('sand', 'other')
df['Description'] = df['Description'].str.replace('dust', 'other')
df['Description'] = df['Description'].str.replace('smoke', 'other')
#standarizing intensity values
df['Description'] = df['Description'].str.replace('few', 'light ')
df['Description'] = df['Description'].str.replace('broken', 'moderate ')
df['Description'] = df['Description'].str.replace('scattered', 'moderate ')
df['Description'] = df['Description'].str.replace('overcast ', 'heavy ')
#multi-categorical dummification
tags = ['clouds','rain','mist','snow','shower','thunderstorm','fog','other']
for i in range(len(tags)):
df[tags[i]]=0
df.loc[df['Description'].str.contains(pat=tags[i])==True, tags[i]] = 1
#creating a weather intensity column
intensity_values=['sky is clear','light','moderate','heavy']
for i in range(len(intensity_values)):
df.loc[df['Description'].str.contains(pat =intensity_values[i])==True, 'Intensity'] = i
#fill in the blanks with moderate
df['Intensity']=df['Intensity'].fillna(2)
return df
def feature_engineer(df): #implementing Hanna's features in a function
df['datetime'] = pd.to_datetime(df['datetime'])
df['year'] = df['datetime'].dt.year
df['month'] = df['datetime'].dt.month
df['day'] = df['datetime'].dt.day
df['hour'] = df['datetime'].dt.hour
df['minute'] = df['datetime'].dt.minute
df['weekday'] = df['datetime'].dt.weekday
df['week'] = df['datetime'].dt.weekofyear
df['quarter'] = df['datetime'].dt.quarter
df['month start'] = df['datetime'].dt.is_month_start
df['month end'] = df['datetime'].dt.is_month_end
df['quarter start'] = df['datetime'].dt.is_quarter_start
df['quarter end'] = df['datetime'].dt.is_quarter_end
#lag features from Duncan
df = df.set_index('datetime')
df['lag1'] = df['Temperature'].shift(periods = 1, fill_value = 0)
df['lag2'] = df['Temperature'].shift(periods = 2, fill_value = 0)
df['lag12'] = df['Temperature'].shift(periods = 12, fill_value = 0)
df['lag30'] = df['Temperature'].shift(periods = 30, fill_value = 0)
#we have to eliminate all those with 0s
df = df[df['lag30'] != 0]
df['max daily temp']=df.resample('D')['Temperature'].transform('max')
df['max daily temp']=df['max daily temp'].shift(24)
df['max daily hum']=df.resample('D')['Humidity'].transform('max')
df['max daily hum']=df['max daily hum'].shift(24)
df['max daily wind speed']=df.resample('D')['Wind Speed'].transform('max')
df['max daily wind speed']=df['max daily wind speed'].shift(24)
df['max daily wind direction']=df.resample('D')['Wind Direction'].transform('max')
df['max daily wind direction']=df['max daily wind direction'].shift(24)
df['max daily pressure']=df.resample('D')['Pressure'].transform('max')
df['max daily pressure']=df['max daily pressure'].shift(24)
df['max weekly temp']=df.resample('W')['Temperature'].transform('max')
df['max weekly temp']=df['max weekly temp'].shift(168)
df['max weekly hum']=df.resample('W')['Humidity'].transform('max')
df['max weekly hum']=df['max weekly hum'].shift(168)
df['max weekly wind speed']=df.resample('W')['Wind Speed'].transform('max')
df['max weekly wind speed']=df['max weekly wind speed'].shift(168)
df['max weekly wind direction']=df.resample('W')['Wind Direction'].transform('max')
df['max weekly wind direction']=df['max weekly wind direction'].shift(168)
df['max weekly pressure']=df.resample('W')['Pressure'].transform('max')
df['max weekly pressure']=df['max weekly pressure'].shift(168)
df['min daily temp']=df.resample('D')['Temperature'].transform('min')
df['min daily temp']=df['min daily temp'].shift(24)
df['min daily hum']=df.resample('D')['Humidity'].transform('min')
df['min daily hum']=df['min daily temp'].shift(24)
df['min daily wind speed']=df.resample('D')['Wind Speed'].transform('min')
df['min daily wind speed']=df['min daily temp'].shift(24)
df['min daily wind direction']=df.resample('D')['Wind Direction'].transform('min')
df['min daily wind direction']=df['min daily temp'].shift(24)
df['min daily pressure']=df.resample('D')['Pressure'].transform('min')
df['min daily pressure']=df['min daily temp'].shift(24)
df['min weekly temp']=df.resample('W')['Temperature'].transform('min')
df['min weekly temp']=df['min weekly temp'].shift(168)
df['min weekly hum']=df.resample('W')['Humidity'].transform('min')
df['min weekly hum']=df['min weekly hum'].shift(168)
df['min weekly wind speed']=df.resample('W')['Wind Speed'].transform('min')
df['min weekly wind speed']=df['min weekly wind speed'].shift(168)
df['min weekly wind direction']=df.resample('W')['Wind Direction'].transform('min')
df['min weekly wind direction']=df['min weekly wind direction'].shift(168)
df['min weekly pressure']=df.resample('W')['Pressure'].transform('min')
df['min weekly pressure']=df['min weekly pressure'].shift(168)
df['mean daily temp']=df.resample('D')['Temperature'].transform('max')
df['mean daily temp']=df['mean daily temp'].shift(24)
df['mean daily hum']=df.resample('D')['Humidity'].transform('max')
df['mean daily hum']=df['mean daily hum'].shift(24)
df['mean daily wind speed']=df.resample('D')['Wind Speed'].transform('max')
df['mean daily wind speed']=df['mean daily wind speed'].shift(24)
df['mean daily wind direction']=df.resample('D')['Wind Direction'].transform('max')
df['mean daily wind direction']=df['mean daily wind direction'].shift(24)
df['mean daily pressure']=df.resample('D')['Pressure'].transform('max')
df['mean daily pressure']=df['mean daily pressure'].shift(24)
df['mean weekly temp']=df.resample('W')['Temperature'].transform('mean')
df['mean weekly temp']=df['mean weekly temp'].shift(168)
df['mean weekly hum']=df.resample('W')['Humidity'].transform('mean')
df['mean weekly hum']=df['mean weekly hum'].shift(168)
df['mean weekly wind speed']=df.resample('W')['Wind Speed'].transform('mean')
df['mean weekly wind speed']=df['mean weekly wind speed'].shift(168)
df['mean weekly wind direction']=df.resample('W')['Wind Direction'].transform('mean')
df['mean weekly wind direction']=df['mean weekly wind direction'].shift(168)
df['mean weekly pressure']=df.resample('W')['Pressure'].transform('mean')
df['mean weekly pressure']=df['mean weekly pressure'].shift(168)
df['rolling_mean_temp'] = df['Temperature'].rolling(window=24).mean()
df['rolling_mean_pressure'] = df['Pressure'].rolling(window=24).mean()
df['rolling_mean_wind_dir'] = df['Wind Direction'].rolling(window=24).mean()
df['rolling_mean_wind_speed'] = df['Wind Speed'].rolling(window=24).mean()
df['rolling_mean_humidity'] = df['Humidity'].rolling(window=24).mean()
df['rolling_min_temp'] = df['Temperature'].rolling(window=24).min()
df['rolling_min_pressure'] = df['Pressure'].rolling(window=24).min()
df['rolling_min_wind_dir'] = df['Wind Direction'].rolling(window=24).min()
df['rolling_min_wind_speed'] = df['Wind Speed'].rolling(window=24).min()
df['rolling_min_humidity'] = df['Humidity'].rolling(window=24).min()
df['rolling_max_temp'] = df['Temperature'].rolling(window=24).max()
df['rolling_max_pressure'] = df['Pressure'].rolling(window=24).max()
df['rolling_max_wind_dir'] = df['Wind Direction'].rolling(window=24).max()
df['rolling_max_wind_speed'] = df['Wind Speed'].rolling(window=24).max()
df['rolling_max_humidity'] = df['Humidity'].rolling(window=24).max()
df['rolling_mean_temp'] = df['Temperature'].expanding(2).mean()
df['rolling_mean_pressure'] = df['Pressure'].expanding(2).mean()
df['rolling_mean_wind_dir'] = df['Wind Direction'].expanding(2).mean()
df['rolling_mean_wind_speed'] = df['Wind Speed'].expanding(2).mean()
df['rolling_mean_humidity'] = df['Humidity'].expanding(2).mean()
df['rolling_min_temp'] = df['Temperature'].expanding(2).min()
df['rolling_min_pressure'] = df['Pressure'].expanding(2).min()
df['rolling_min_wind_dir'] = df['Wind Direction'].expanding(2).min()
df['rolling_min_wind_speed'] = df['Wind Speed'].expanding(2).min()
df['rolling_min_humidity'] = df['Humidity'].expanding(2).min()
df['rolling_max_temp'] = df['Temperature'].expanding(2).max()
df['rolling_max_pressure'] = df['Pressure'].expanding(2).max()
df['rolling_max_wind_dir'] = df['Wind Direction'].expanding(2).max()
df['rolling_max_wind_speed'] = df['Wind Speed'].expanding(2).max()
df['rolling_max_humidity'] = df['Humidity'].expanding(2).max()
return(df)
def feature_engineer_important(df):
df = feature_engineer(df) #create new features using above
df = df.dropna()
# split into input and output
X = df.drop('Temperature',axis=1)
X = X.drop('Description',axis=1)
y = df['Temperature']
# perform feature selection
rfe = RFE(DecisionTreeRegressor(random_state=1), n_features_to_select=20)
fit = rfe.fit(X, y)
important_features = []
names = X.columns
for i in range(len(fit.support_)):
if fit.support_[i]:
important_features.append((names[i]))
df = df[important_features+['Temperature']] #only return the top 15 most important features
return(df)
def split(sequence, n_timestamp):
X, y = [], []
for i in range(len(sequence)):
end = i + n_timestamp
if end > len(sequence)-1:
break
sequence_x, sequence_y = sequence[i:end], sequence[end]
X.append(sequence_x)
y.append(sequence_y)
return np.array(X), np.array(y)
def split_multiple(sequence, n_timestamp, target):
X, y = [], []
for i in range(len(sequence)):
end = i + n_timestamp
if end > len(sequence)-1:
break
sequence_x, sequence_y = sequence[i:end], target[end]
X.append(sequence_x)
y.append(sequence_y)
return np.array(X), np.array(y)
def seasons(df): #creates the seasons columns
df['datetime']= pd.to_datetime(df['datetime'])
df['season'] = (df['datetime'].dt.month%12 + 3)//3
seasons = {1: 'Winter',2: 'Spring',3: 'Summer',4: 'Autumn'}
df['season_name'] = df['season'].map(seasons)
df = df.drop(['season'],axis=1)
return df
def cont_into_cat(df): #makes the continuous variables into categoricals based on quantiles
labels = ['low', 'medium', 'high']
df['Wind_Speed_Quantiles'] = pd.qcut(df['Wind Speed'], q=3,labels=labels,precision=0)
df['Humidity_Quantiles'] = pd.qcut(df['Humidity'], q=3,labels=labels,precision=0)
df['Pressure_Quantiles'] = pd.qcut(df['Pressure'], q=3,labels=labels,precision=0)
return df