-
Notifications
You must be signed in to change notification settings - Fork 0
/
worker.py
504 lines (484 loc) · 28.3 KB
/
worker.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
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
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
import numpy as np
import pandas as pd
import s3fs
import sys
import os
import pickle
import tempfile
from datetime import datetime
# THEFT, BURGLARY, MOTOR VEHICLE THEFT, ROBBERY THEFT 1
# BATTERY, CRIM SEXUAL ASSAULT, SEX OFFENSE SEXUAL ASSAULT 2
# NARCOTICS, OTHER NARCOTIC VIOLATION NARCOTICS 3
# ASSAULT, INTIMIDATION ASSAULT 4
# OTHER OFFENSE OTHER OFFENSE 5
# DECEPTIVE PRACTICE DECEPTIVE PRACTICE 6
# CRIMINAL TRESPASS CRIMINAL TRESPASS 7
# WEAPONS VIOLATION, CONCEALED CARRY LICENSE VIOLATION WEAPONS VIOLATION 8
# PUBLIC INDECENCY, PUBLIC PEACE VIOLATION PUBLIC INDECENCY 9
# OFFENSE INVOLVING CHILDREN OFFENSE INVOLVING CHILDREN 10
# PROSTITUTION PROSTITUTION 11
# INTERFERENCE WITH PUBLIC OFFICER INTERFERENCE WITH PUBLIC OFFICER 12
# HOMICIDE HOMICIDE 13
# ARSON, CRIMINAL DAMAGE ARSON 14
# GAMBLING GAMBLING 15
# LIQUOR LAW VIOLATION LIQUOR LAW VIOLATION 16
# KIDNAPPING KIDNAPPING 17
# STALKING, OBSCENITY STALKING 18
# NON - CRIMINAL, NON-CRIMINAL (SUBJECT SPECIFIED) NON - CRIMINAL 19
# HUMAN TRAFFICKING HUMAN TRAFFICKING 20
# RITUALISM RITUALISM 21
# DOMESTIC VIOLENCE DOMESTIC VIOLENCE 22
def round_hour(dt):
if (dt.hour >= 0) and (dt.hour <= 6):
return datetime(dt.year, dt.month, dt.day, 6,0)
elif (dt.hour > 6) and (dt.hour <= 11):
return datetime(dt.year, dt.month, dt.day, 11,0)
elif (dt.hour > 11) and (dt.hour <= 17):
return datetime(dt.year, dt.month, dt.day, 17,0)
else:
return datetime(dt.year, dt.month, dt.day, 23,0)
community_to_code = { 'Rogers Park':1,
'West Ridge':2,
'Uptown':3,
'Lincoln Square':4,
'North Center':5,
'Lake View':6,
'Lincoln Park':7,
'Near North Side':8,
'Edison Park':9,
'Norwood Park':10,
'Jefferson Park':11,
'Forest Glen':12,
'North Park':13,
'Albany Park':14,
'Portage Park':15,
'Irving Park':16,
'Dunning':17,
'Montclare':18,
'Belmont Cragin':19,
'Hermosa':20,
'Avondale':21,
'Logan Square':22,
'Humboldt Park':23,
'West Town':24,
'Austin':25,
'West Garfield Park':26,
'East Garfield Park':27,
'Near West Side':28,
'North Lawndale':29,
'South Lawndale':30,
'Lower West Side':31,
'The Loop':32,
'Near South Side':33,
'Armour Square':34,
'Douglas':35,
'Oakland':36,
'Fuller Park':37,
'Grand Boulevard':38,
'Kenwood':39,
'Washington Park':40,
'Hyde Park':41,
'Woodlawn':42,
'South Shore':43,
'Chatham':44,
'Avalon Park':45,
'South Chicago':46,
'Burnside':47,
'Calumet Heights':48,
'Roseland':49,
'Pullman':50,
'South Deering':51,
'East Side':52,
'West Pullman':53,
'Riverdale':54,
'Hegewisch':55,
'Garfield Ridge':56,
'Archer Heights':57,
'Brighton Park':58,
'McKinley Park':59,
'Bridgeport':60,
'New City':61,
'West Elsdon':62,
'Gage Park':63,
'Clearing':64,
'West Lawn':65,
'Chicago Lawn':66,
'West Englewood':67,
'Englewood':68,
'Greater Grand Crossing':69,
'Ashburn':70,
'Auburn Gresham':71,
'Beverly':72,
'Washington Heights':73,
'Mount Greenwood':74,
'Morgan Park':75,
'O\'Hare':76,
'Edgewater':77
}
crime_classes = {1:'THEFT', 2:'SEXUAL ASSAULT', 3:'NARCOTICS', 4:'ASSAULT', 5:'OTHER OFFENSE', 6:'DECEPTIVE PRACTICE',
7:'CRIMINAL TRESPASS', 8:'WEAPONS VIOLATION', 9:'PUBLIC INDECENCY', 10:'OFFENSE INVOLVING CHILDREN',
11:'PROSTITUTION', 12:'INTERFERENCE WITH PUBLIC OFFICER', 13:'HOMICIDE', 14:'ARSON', 15:'GAMBLING',
16:'LIQUOR LAW VIOLATION', 17:'KIDNAPPING', 18:'STALKING', 19:'NON - CRIMINAL', 20:'HUMAN TRAFFICKING',
21:'RITUALISM',22:'DOMESTIC VIOLENCE'}
print('[' + str(datetime.now()) + '] Reading Crimes dataset...')
sys.stdout.flush()
s3fs.S3FileSystem.read_timeout = 5184000 # one day
s3fs.S3FileSystem.connect_timeout = 5184000 # one day
try:
#file = './data/Crimes_-_2001_to_present.csv' # This line to read from local disk
#file = 's3://w210policedata/datasets/Crimes_test.csv' # This line to read quick test file from S3
file = 's3://w210policedata/datasets/Crimes_-_2001_to_present.csv' # This line to read from S3
crimes = pd.read_csv(file,sep=',', error_bad_lines=False, index_col='Date', dtype='unicode')
except Exception as e:
print('[' + str(datetime.now()) + '] Error reading input dataset: '+file)
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
print('[' + str(datetime.now()) + '] Processing and transforming data...')
sys.stdout.flush()
try:
crimes = crimes.iloc[:, 3: ]
print('[' + str(datetime.now()) + '] * Grouping similar crime types...')
sys.stdout.flush()
crimes = crimes.replace(['THEFT', 'BURGLARY', 'MOTOR VEHICLE THEFT', 'ROBBERY' ,'BATTERY', 'CRIM SEXUAL ASSAULT',
'SEX OFFENSE' , 'NARCOTICS','OTHER NARCOTIC VIOLATION' , 'ASSAULT', 'INTIMIDATION' ,
'OTHER OFFENSE' , 'DECEPTIVE PRACTICE' , 'CRIMINAL TRESPASS' , 'WEAPONS VIOLATION' ,
'CONCEALED CARRY LICENSE VIOLATION','PUBLIC INDECENCY', 'PUBLIC PEACE VIOLATION',
'OFFENSE INVOLVING CHILDREN','PROSTITUTION','INTERFERENCE WITH PUBLIC OFFICER','HOMICIDE',
'ARSON', 'CRIMINAL DAMAGE','GAMBLING','LIQUOR LAW VIOLATION','KIDNAPPING','STALKING',
'OBSCENITY','NON - CRIMINAL','NON-CRIMINAL', 'NON-CRIMINAL (SUBJECT SPECIFIED)','HUMAN TRAFFICKING',
'RITUALISM','DOMESTIC VIOLENCE']
,[crime_classes[1],crime_classes[1],crime_classes[1],crime_classes[1],
crime_classes[2],crime_classes[2],crime_classes[2],
crime_classes[3],crime_classes[3],
crime_classes[4],crime_classes[4],
crime_classes[5],
crime_classes[6],
crime_classes[7],
crime_classes[8],crime_classes[8],
crime_classes[9],crime_classes[9],
crime_classes[10],
crime_classes[11],
crime_classes[12],
crime_classes[13],
crime_classes[14],crime_classes[14],
crime_classes[15],
crime_classes[16],
crime_classes[17],
crime_classes[18],crime_classes[18],
crime_classes[19],crime_classes[19],crime_classes[19],
crime_classes[17],
crime_classes[5],
crime_classes[4]])
print('[' + str(datetime.now()) + '] * Filtering columns...')
sys.stdout.flush()
crimes = crimes[['Primary Type','Community Area']]
print('[' + str(datetime.now()) + '] * Removing problem lines...')
sys.stdout.flush()
crimes = crimes.dropna(axis=0,how='any')
print('[' + str(datetime.now()) + '] * Converting time format...')
sys.stdout.flush()
crimes.index = pd.to_datetime(crimes.index)
except Exception as e:
print('[' + str(datetime.now()) + '] Error performing first part of transformations.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
# print('[' + str(datetime.now()) + '] Performing transformations for time series models...')
# sys.stdout.flush()
# try:
# print('[' + str(datetime.now()) + '] * Convert date and time to blocks of hours...')
# sys.stdout.flush()
# crimes_ts = crimes.reset_index()
# crimes_ts['Date'] = crimes_ts['Date'].apply(round_hour)
# print('[' + str(datetime.now()) + '] * Consolidating data and filling empty rows...')
# sys.stdout.flush()
# crimes_ts = crimes_ts.groupby(['Date','Primary Type','Community Area'])\
# .size().unstack().fillna(0).stack().reset_index(name='counts')
# print('[' + str(datetime.now()) + '] * One-Hot Encoding categorical variables...')
# sys.stdout.flush()
# crimes_ts = pd.concat([crimes_ts,pd.get_dummies(crimes_ts['Primary Type'], prefix='primaryType')],axis=1)
# print('[' + str(datetime.now()) + '] - Primary Type complete')
# sys.stdout.flush()
# crimes_ts = pd.concat([crimes_ts,pd.get_dummies(crimes_ts['Community Area'], prefix='communityArea')],axis=1)
# print('[' + str(datetime.now()) + '] - Community Area complete')
# sys.stdout.flush()
# print('[' + str(datetime.now()) + '] * Sorting data frame to group communities and crime types...')
# sys.stdout.flush()
# crimes_ts.sort_values(by=['Community Area', 'Primary Type', 'Date'], inplace=True)
# print('[' + str(datetime.now()) + '] * Dropping unused columns...')
# sys.stdout.flush()
# comms = crimes_ts['Community Area'].unique()
# cts = crimes_ts['Primary Type'].unique()
# crimes_ts.drop(columns=['Date','Primary Type','Community Area'], inplace=True)
# print('[' + str(datetime.now()) + '] * Sorting column order...')
# sys.stdout.flush()
# cols = crimes_ts.columns.tolist()[1:] + crimes_ts.columns.tolist()[:1]
# crimes_ts = crimes_ts[cols]
# print('[' + str(datetime.now()) + '] * Setting dataframe to sparse format...')
# sys.stdout.flush()
# crimes_ts = crimes_ts.to_sparse(fill_value=0)
# print('[' + str(datetime.now()) + '] * Re-strutucturing data into time lag format...')
# sys.stdout.flush()
# lag = 4*7*16 # 16 weeks
# val_size = 240 # 240 time periods per community per type of crime (60 days)
# partials_train = []
# partials_val = []
# comm_count = 1
# for comm in comms:
# print('[' + str(datetime.now()) + '] - Running community '+str(comm_count)+' of '+str(len(comms)))
# sys.stdout.flush()
# ct_count = 1
# for ct in cts:
# print('[' + str(datetime.now()) + '] # Running primary type '+ct+' (' + str(ct_count) + ' of ' + str(len(cts))+ ')' )
# sys.stdout.flush()
# crimes_ts_pt = crimes_ts[((crimes_ts['communityArea_'+comm] == 1) & (crimes_ts['primaryType_'+ct] == 1))]
# print('[' + str(datetime.now()) + '] > Dataset sliced' )
# sys.stdout.flush()
# columns = [crimes_ts_pt.shift(i) for i in range(1, lag+1)]
# print('[' + str(datetime.now()) + '] > Lagged time columns created' )
# sys.stdout.flush()
# columns.append(crimes_ts_pt.iloc[:,-1:])
# crimes_ts_pt = pd.concat(columns, axis=1)
# print('[' + str(datetime.now()) + '] > Columns concatenated' )
# sys.stdout.flush()
# crimes_ts_pt = crimes_ts_pt[lag:]
# print('[' + str(datetime.now()) + '] > Removed initial rows' )
# sys.stdout.flush()
# partials_train.append(crimes_ts_pt[val_size+1:])
# partials_val.append(crimes_ts_pt[:-val_size])
# print('[' + str(datetime.now()) + '] > Done' )
# sys.stdout.flush()
# ct_count = ct_count + 1
# comm_count = comm_count+1
# crimes_ts_train = pd.concat(partials_train,ignore_index=True)
# crimes_ts_val = pd.concat(partials_val,ignore_index=True)
# except Exception as e:
# print('[' + str(datetime.now()) + '] Error performing transformations for time-series.')
# print(e)
# print('[' + str(datetime.now()) + '] Aborting...')
# sys.stdout.flush()
# sys.exit(1)
#
# print('[' + str(datetime.now()) + '] Writing time-series datasets...')
# sys.stdout.flush()
# try:
# #output = './data/ProcessedDataset.parquet' # This line to write to local disk
# output_train = 's3://w210policedata/datasets/OneHotEncodedTSDatasetTRAIN.parquet' # This line to write to S3
# output_val = 's3://w210policedata/datasets/OneHotEncodedTSDatasetVAL.parquet'
# crimes_ts_train.to_parquet(output_train,index=False)
# crimes_ts_val.to_parquet(output_val,index=False)
# del columns
# del crimes_ts
# del crimes_ts_pt
# del crimes_ts_train
# del crimes_ts_val
# del partials_train
# del partials_val
# except:
# print('[' + str(datetime.now()) + '] Error writing time-series output dataset: '+output)
# print('[' + str(datetime.now()) + '] Aborting...')
# sys.exit(1)
print('[' + str(datetime.now()) + '] Continuing data transformation for fixed-time models...')
sys.stdout.flush()
try:
crimes = crimes.reset_index()
print('[' + str(datetime.now()) + '] * Creating new features from columns...')
sys.stdout.flush()
crimes['Weekday'] = crimes['Date'].dt.dayofweek
crimes['Week of Year'] = crimes['Date'].dt.weekofyear
crimes['Hour of the Day'] = crimes['Date'].dt.hour
crimes['Hour of the Day'].replace([0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23],
['DAWN','DAWN','DAWN','DAWN','DAWN','DAWN','DAWN',
'MORNING','MORNING','MORNING','MORNING','MORNING',
'AFTERNOON','AFTERNOON','AFTERNOON','AFTERNOON','AFTERNOON','AFTERNOON',
'EVENING','EVENING','EVENING','EVENING','EVENING','EVENING'],
inplace=True)
crimes['Year'] = pd.to_datetime(crimes['Date']).dt.year
print('[' + str(datetime.now()) + '] * Consolidating data and filling empty rows...')
sys.stdout.flush()
crimes = crimes.groupby(['Year','Primary Type','Community Area','Weekday','Week of Year','Hour of the Day'])\
.size().unstack().fillna(0).stack().reset_index(name='counts')
except Exception as e:
print('[' + str(datetime.now()) + '] Error performing second part of transformations.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
try:
print('[' + str(datetime.now()) + '] Reading CMAP dataset for additional features...')
sys.stdout.flush()
file = 's3://w210policedata/datasets/CMAP_dataset.csv' # This line to read from S3
cmap = pd.read_csv(file,sep=',', error_bad_lines=False, dtype='unicode')
except Exception as e:
print('[' + str(datetime.now()) + '] Error reading CMAP dataset: '+file)
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
try:
print('[' + str(datetime.now()) + '] Preparing and transforming CMAP dataset...')
print('[' + str(datetime.now()) + '] * Replacing community names with community codes...')
sys.stdout.flush()
cmap['Community Area'] = cmap['GEOG'].map(community_to_code)
cmap['Community Area'] = cmap['Community Area'].map(str)
print('[' + str(datetime.now()) + '] * Transforming columns...')
sys.stdout.flush()
cmap['socioEconomic_medianAge'] = pd.to_numeric(cmap['MED_AGE'])
cmap['socioEconomic_medianIncome'] = pd.to_numeric(cmap['MEDINC'])
cmap['socioEconomic_popInHouseholds'] = pd.to_numeric(cmap['POP_HH'])/pd.to_numeric(cmap['TOT_POP'])
cmap['schooling_lessHighSchool'] = pd.to_numeric(cmap['LT_HS'])/pd.to_numeric(cmap['POP_25OV'])
cmap['schooling_highSchool'] = pd.to_numeric(cmap['HS'])/pd.to_numeric(cmap['POP_25OV'])
cmap['schooling_someCollege'] = pd.to_numeric(cmap['SOME_COLL'])/pd.to_numeric(cmap['POP_25OV'])
cmap['schooling_Associate'] = pd.to_numeric(cmap['ASSOC'])/pd.to_numeric(cmap['POP_25OV'])
cmap['schooling_Bachelor'] = pd.to_numeric(cmap['BACH'])/pd.to_numeric(cmap['POP_25OV'])
cmap['schooling_Graduate'] = pd.to_numeric(cmap['GRAD_PROF'])/pd.to_numeric(cmap['POP_25OV'])
cmap['housing_Occupied'] = pd.to_numeric(cmap['TOT_HH'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_Vacant'] = pd.to_numeric(cmap['VAC_HU'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_OwnerOccupied'] = pd.to_numeric(cmap['OWN_OCC_HU'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_RenterOccupied'] = pd.to_numeric(cmap['RENT_OCC_HU'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_SingleFamilyDetached'] = pd.to_numeric(cmap['HU_SNG_DET'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_SingleFamilyAttached'] = pd.to_numeric(cmap['HU_SNG_ATT'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_TwoUnits'] = pd.to_numeric(cmap['HU_2UN'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_ThreeOrFourUnits'] = pd.to_numeric(cmap['HU_3_4UN'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_FiveOrMoreUnits'] = pd.to_numeric(cmap['HU_GT_5UN'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_medianNumberRooms'] = pd.to_numeric(cmap['MED_ROOMS'])
cmap['housing_Later2000'] = pd.to_numeric(cmap['HA_AFT2000'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_1970to1999'] = pd.to_numeric(cmap['HA_70_00'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_1940to1969'] = pd.to_numeric(cmap['HA_40_70'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_Before1940'] = pd.to_numeric(cmap['HA_BEF1940'])/pd.to_numeric(cmap['HU_TOT'])
cmap['housing_medianHouseAge'] = pd.to_numeric(cmap['MED_HA'])
cmap['socioEconomic_inLaborForce'] = pd.to_numeric(cmap['IN_LBFRC'])/pd.to_numeric(cmap['POP_16OV'])
cmap['socioEconomic_employed'] = pd.to_numeric(cmap['EMP'])/pd.to_numeric(cmap['IN_LBFRC'])
cmap['socioEconomic_unemployed'] = pd.to_numeric(cmap['UNEMP'])/pd.to_numeric(cmap['IN_LBFRC'])
cmap['socioEconomic_notInLaborForce'] = pd.to_numeric(cmap['NOT_IN_LBFRC'])/pd.to_numeric(cmap['POP_16OV'])
cmap['commute_carAlone'] = pd.to_numeric(cmap['DROVE_AL'])/pd.to_numeric(cmap['TOT_COMM'])
cmap['commute_carpool'] = pd.to_numeric(cmap['CARPOOL'])/pd.to_numeric(cmap['TOT_COMM'])
cmap['commute_transit'] = pd.to_numeric(cmap['TRANSIT'])/pd.to_numeric(cmap['TOT_COMM'])
cmap['commute_walkOrBike'] = pd.to_numeric(cmap['WALK_BIKE'])/pd.to_numeric(cmap['TOT_COMM'])
cmap['commute_other'] = pd.to_numeric(cmap['COMM_OTHER'])/pd.to_numeric(cmap['TOT_COMM'])
cmap['commute_averageVehicleMilesTravelled'] = pd.to_numeric(cmap['AVG_VMT'])
cmap['socioEconomic_noVehiclesAvailable'] = pd.to_numeric(cmap['NO_VEH'])/pd.to_numeric(cmap['TOT_HH'])
cmap['socioEconomic_oneVehicleAvailable'] = pd.to_numeric(cmap['ONE_VEH'])/pd.to_numeric(cmap['TOT_HH'])
cmap['socioEconomic_twoVehiclesAvailable'] = pd.to_numeric(cmap['TWO_VEH'])/pd.to_numeric(cmap['TOT_HH'])
cmap['socioEconomic_threeOrMoreVehiclesAvailable'] = pd.to_numeric(cmap['THREEOM_VEH'])/pd.to_numeric(cmap['TOT_HH'])
cmap['lifeQuality_accessibleParkAcreage'] = pd.to_numeric(cmap['OPEN_SPACE_PER_1000'])
cmap['landUse_singleFamilyResidential'] = pd.to_numeric(cmap['Sfperc'])
cmap['landUse_multiFamilyResidential'] = pd.to_numeric(cmap['Mfperc'])
cmap['landUse_commercial'] = pd.to_numeric(cmap['COMMperc'])
cmap['landUse_industrial'] = pd.to_numeric(cmap['INDperc'])
cmap['landUse_institutional'] = pd.to_numeric(cmap['INSTperc'])
cmap['landUse_mixedUse'] = pd.to_numeric(cmap['MIXperc'])
cmap['landUse_transportation'] = pd.to_numeric(cmap['TRANSperc'])
cmap['landUse_agricultural'] = pd.to_numeric(cmap['Agperc'])
cmap['landUse_openSpace'] = pd.to_numeric(cmap['OPENperc'])
cmap['landUse_vacant'] = pd.to_numeric(cmap['VACperc'])
print('[' + str(datetime.now()) + '] * Filtering columns...')
sys.stdout.flush()
regex="(Community Area)|(socioEconomic_)|(schooling_)|(housing_)|(commute_)|(lifeQuality_)|(landUse_)"
cmap = cmap.filter(regex=regex,axis=1)
print('[' + str(datetime.now()) + '] * Filling in NAs...')
sys.stdout.flush()
cmap.fillna(0, inplace=True)
except Exception as e:
print('[' + str(datetime.now()) + '] Error transforming CMAP dataset.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
try:
print('[' + str(datetime.now()) + '] Writing additional features dataset...')
sys.stdout.flush()
# output = 's3://w210policedata/datasets/AdditionalFeatures.parquet' # This line to write to S3
# cmap.to_parquet(output,index=False)
output = 's3://w210policedata/datasets/AdditionalFeatures.csv' # This line to write to S3
cmap.to_csv(output,index=False)
except Exception as e:
print('[' + str(datetime.now()) + '] Error writing additional features dataset.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
try:
print('[' + str(datetime.now()) + '] Combining Additional Features and Crimes datasets...')
sys.stdout.flush()
crimes = pd.merge(crimes, cmap, on='Community Area')
except Exception as e:
print('[' + str(datetime.now()) + '] Error combining additional features and crimes datasets.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
print('[' + str(datetime.now()) + '] Writing intermediate dataset...')
sys.stdout.flush()
try:
#output = './data/ProcessedDataset.parquet' # This line to write to local disk
output = 's3://w210policedata/datasets/ProcessedDataset.parquet' # This line to write to S3
crimes.to_parquet(output,index=False)
except:
print('[' + str(datetime.now()) + '] Error writing intermediate output dataset: '+output)
print('[' + str(datetime.now()) + '] Aborting...')
sys.exit(1)
try:
print('[' + str(datetime.now()) + '] Continuing data processing and transformation...')
sys.stdout.flush()
print('[' + str(datetime.now()) + '] * One-Hot Encoding categorical variables...')
sys.stdout.flush()
crimes = pd.concat([crimes,pd.get_dummies(crimes['Primary Type'], prefix='primaryType')],axis=1)
print('[' + str(datetime.now()) + '] - Primary Type complete')
sys.stdout.flush()
crimes = pd.concat([crimes,pd.get_dummies(crimes['Community Area'], prefix='communityArea')],axis=1)
print('[' + str(datetime.now()) + '] - Community Area complete')
sys.stdout.flush()
crimes = pd.concat([crimes,pd.get_dummies(crimes['Weekday'], prefix='weekDay')],axis=1)
print('[' + str(datetime.now()) + '] - Weekday complete')
sys.stdout.flush()
crimes = pd.concat([crimes,pd.get_dummies(crimes['Week of Year'], prefix='weekYear')],axis=1)
print('[' + str(datetime.now()) + '] - Week of year complete')
sys.stdout.flush()
crimes = pd.concat([crimes,pd.get_dummies(crimes['Hour of the Day'], prefix='hourDay')],axis=1)
print('[' + str(datetime.now()) + '] - Hour of day complete')
sys.stdout.flush()
print('[' + str(datetime.now()) + '] * Dropping unused columns...')
sys.stdout.flush()
crimes.drop(columns=['Year','Primary Type','Community Area','Weekday','Week of Year','Hour of the Day'], inplace=True)
except Exception as e:
print('[' + str(datetime.now()) + '] Error performing second part of transformations.')
print('[' + str(datetime.now()) + '] Aborting...')
sys.stdout.flush()
sys.exit(1)
print('[' + str(datetime.now()) + '] Writing one-hot encoded dataset...')
sys.stdout.flush()
try:
#output = './data/OneHotEncodedDataset.parquet' # This line to write to local disk
output = 's3://w210policedata/datasets/OneHotEncodedDataset.parquet' # This line to write to S3
crimes.to_parquet(output,index=False)
except:
print('[' + str(datetime.now()) + '] Error writing output dataset: '+output)
print('[' + str(datetime.now()) + '] Aborting...')
sys.exit(1)
print('[' + str(datetime.now()) + '] Building available features file...')
sys.stdout.flush()
features = [
{'feature': 'Community Area', 'column': 'communityArea', 'onehot-encoded': True, 'ethnically_biased': True, 'optional': True},
{'feature': 'Crime Type', 'column': 'primaryType', 'onehot-encoded': True, 'ethnically_biased': False, 'optional': False},
{'feature': 'Day of the Week', 'column': 'weekDay', 'onehot-encoded': True, 'ethnically_biased': False, 'optional': False},
{'feature': 'Week of the Year', 'column': 'weekYear', 'onehot-encoded': True, 'ethnically_biased': False, 'optional': False},
{'feature': 'Period of the Day', 'column': 'hourDay', 'onehot-encoded': True, 'ethnically_biased': False, 'optional': False},
{'feature': 'Socioeconomic Data', 'column': 'socioEconomic', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True},
{'feature': 'Schooling', 'column': 'schooling', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True},
{'feature': 'Housing', 'column': 'housing', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True},
{'feature': 'Commute', 'column': 'commute', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True},
{'feature': 'Quality of Life', 'column': 'lifeQuality', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True},
{'feature': 'Land Use', 'column': 'landUse', 'onehot-encoded': False, 'ethnically_biased': False, 'optional': True}
]
print('[' + str(datetime.now()) + '] Writing available features file...')
sys.stdout.flush()
try:
#output = './data/OneHotEncodedDataset.parquet' # This line to write to local disk
output = 'w210policedata/datasets/AvailableFeatures.pickle' # This line to write to S3
s3 = s3fs.S3FileSystem(anon=False)
with s3.open(output, "wb") as json_file:
pickle.dump(features, json_file, protocol=pickle.HIGHEST_PROTOCOL)
json_file.close()
except:
print('[' + str(datetime.now()) + '] Error writing features file: '+output)
print('[' + str(datetime.now()) + '] Aborting...')
sys.exit(1)
print('[' + str(datetime.now()) + '] Finished!')
sys.exit(0)