-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathcrime_portal.py
154 lines (134 loc) · 5.73 KB
/
crime_portal.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
"""
CAPP 30254: Final project
This file contains the code used to obtain the data recquired from the Chicago
Open Data Portal.
"""
import pandas as pd
import numpy as np
from sodapy import Socrata
from sklearn.preprocessing import MinMaxScaler
def read_beat_data():
crime_beat_quartiles = pd.read_csv('data/crime_beat_quartiles.csv')
crime_beat_quartiles['crime_month'] = \
pd.to_datetime(crime_beat_quartiles.crime_month)
crime_beat_quartiles['crime_month'] = \
crime_beat_quartiles.crime_month.map(lambda x: x.strftime('%Y-%m'))
beat = pd.read_csv('data/data_area.csv')
beat = beat.loc[beat.area_type == 'beat']
return beat, crime_beat_quartiles
def beat_quartile_trrs(officer_df, trr_df, end_date_set, feat_dict, train=True):
'''
Adds features indicating average number of trrs per year in beats of
the first, second, third, and fourth quartiles of crime by month.
'''
beat, crime_beat_quartiles = read_beat_data()
trr_df['trr_datetime'] = \
trr_df.apply(
lambda x: x['trr_datetime'].tz_localize(None), axis=1)
trr = trr_df.loc[(trr_df.trr_datetime <= end_date_set)]
total_years = (end_date_set - np.datetime64('2010-01-01')) \
/ np.timedelta64(365, 'D')
trr['trr_month'] = trr.trr_datetime.map(lambda x: x.strftime('%Y-%m'))
trr['beat'] = trr.beat.astype('int')
merged_quartiles = trr.merge(
crime_beat_quartiles,
left_on=['beat', 'trr_month'],
right_on=['beat', 'crime_month'])
beat_officer_quartiles = pd.DataFrame(merged_quartiles.groupby(
['beat', 'officer_id', 'quartile'])['id'].nunique()).reset_index()
officer_quartiles = pd.DataFrame(beat_officer_quartiles.groupby(
['officer_id', 'quartile'])['id'].nunique()).reset_index()
officer_quartiles['id'] = officer_quartiles.id.map(
lambda x: x / total_years)
officer_quartiles = officer_quartiles.pivot_table(
'id', 'officer_id', 'quartile').fillna(0)
officer_quartiles.rename(
columns={'first': 'first_quartile_trrs',
'second': 'second_quartile_trrs',
'third': 'third_quartile_trrs',
'fourth': 'fourth_quartile_trrs'}, inplace=True)
officer_df = officer_df.merge(
officer_quartiles, how='left', left_on='id', right_on='officer_id')
newvars = ['first_quartile_trrs', 'second_quartile_trrs',
'third_quartile_trrs', 'fourth_quartile_trrs']
for col_name in newvars:
officer_df[col_name] = officer_df[col_name].fillna(
officer_df[col_name].mean())
if train:
for col in newvars:
scaler = MinMaxScaler()
officer_df[col] = scaler.fit_transform(np.array(officer_df[col])\
.reshape(-1, 1))
feat_dict[col] = scaler
return officer_df, feat_dict, newvars
else:
for col in newvars:
scaler = feat_dict[col]
officer_df[col] = scaler.transform(np.array(officer_df[col])\
.reshape(-1, 1))
return officer_df
def beat_quartile_complaints(officer_df, allegation_df, end_date_set,
feat_dict, train=True):
'''
Adds features indicating average number of complaints per year in beats of
the first, second, third, and fourth quartiles of crime by month.
'''
#end_date_set = end_date_set.tz_localize(None)
'''
allegation_df['incident_date'] = \
allegation_df.apply(
lambda x: x['incident_date'].tz_localize(None), axis=1)
'''
beat, crime_beat_quartiles = read_beat_data()
allegation_df = allegation_df.loc[
pd.notnull(allegation_df.beat_id)]
allegation_df['beat_id'] = \
allegation_df.beat_id.astype('int')
allegation_df = \
allegation_df.merge(
beat, left_on='beat_id', right_on='id')
allegation_df['name'] = \
allegation_df.name.astype('int')
allegation_df = allegation_df.loc[
allegation_df['incident_date'] <= end_date_set]
allegation_df['incident_month'] = \
allegation_df.incident_date.map(lambda x: x.strftime('%Y-%m'))
merged_quartiles = allegation_df.merge(
crime_beat_quartiles,
left_on=['name', 'incident_month'],
right_on=['beat', 'crime_month'])
officer_quartiles = pd.DataFrame(
merged_quartiles.groupby(
['officer_id', 'quartile'])['allegation_id'].nunique()).\
reset_index()
total_years = (end_date_set - np.datetime64('2010-01-01')) / \
np.timedelta64(365, 'D')
officer_quartiles['allegation_id'] = officer_quartiles.allegation_id.map(
lambda x: x / total_years)
officer_quartiles = officer_quartiles.pivot_table(
'allegation_id', 'officer_id', 'quartile').fillna(0)
officer_quartiles.rename(
columns={'first': 'first_quartile',
'second': 'second_quartile',
'third': 'third_quartile',
'fourth': 'fourth_quartile'}, inplace=True)
officer_df = officer_df.merge(
officer_quartiles, how='left', left_on='id', right_on='officer_id')
newvars = ['first_quartile', 'second_quartile', 'third_quartile',
'fourth_quartile']
for col_name in newvars:
officer_df[col_name] = officer_df[col_name].fillna(
officer_df[col_name].mean())
if train:
for col in newvars:
scaler = MinMaxScaler()
officer_df[col] = scaler.fit_transform(np.array(officer_df[col])\
.reshape(-1, 1))
feat_dict[col] = scaler
return officer_df, feat_dict, newvars
else:
for col in newvars:
scaler = feat_dict[col]
officer_df[col] = scaler.transform(np.array(officer_df[col])\
.reshape(-1, 1))
return officer_df