-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathread_incomplete_data.py
174 lines (136 loc) · 7.15 KB
/
read_incomplete_data.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
import pandas as pd
import numpy as np
from pandas.api.types import CategoricalDtype
def load_and_preprocess(trend_name=None):
data = load(trend_name=trend_name)
dataset, attributes, descriptives = define_attributes(data=data, trend_name='MissingDataExperiments', skip_attributes=['Remainer'], outcome_attribute=['Leaver'])
#dataset, descriptives = missing_data_method(dataset=dataset, descriptives=descriptives)
finished_dataset = reset_attribute_type(dataset=dataset, descriptives=descriptives)
return finished_dataset, attributes, descriptives
def load(trend_name=None):
location = 'data_input/Brexit/MissingDataExperiments/' + trend_name
dataset = pd.read_csv(location)
print(dataset.head())
print(dataset.dtypes)
print(dataset.shape)
print(dataset.isnull().sum())
return dataset
def define_attributes(data=None, trend_name=None, skip_attributes=None, outcome_attribute=None):
time_attribute = ['Wave']
#outcome_attribute = ['euspeed1num']
data_sorted = data.sort_values(['Wave'], ascending=[True]).reset_index(drop=True)
data_sorted['id'] = np.arange(len(data_sorted))
id_attribute = ['id']
#print(data_sorted.dtypes)
num_atts = ['age', 'Tradeimmig']
bin_atts = ['sex']
if trend_name in ['Leaver_with', 'Remainer', 'Remainer_plus_Leaver', 'MissingDataExperiments']:
nom_atts = ['Hindsight', 'work_stat', 'work_organisation', 'work_type', 'region', 'EURef2016']
ord_atts = ['Poscountry', 'Posind', 'Govthand', 'profile_gross_personal', 'education_age', 'socialgradeCIE2']
elif trend_name in ['Leaver_without']:
nom_atts = ['work_stat', 'work_organisation', 'work_type', 'region', 'EURef2016']
ord_atts = ['Govthand', 'profile_gross_personal', 'education_age', 'socialgradeCIE2']
skip_attributes_temp = []
if not skip_attributes is None:
for var in skip_attributes:
skip_attributes_temp.append(var)
if var in ord_atts:
ord_atts.remove(var)
skip_attributes = skip_attributes_temp
#print(ord_atts)
#print(skip_attributes)
dataset = data_sorted.drop(skip_attributes, axis=1)
descriptives = {'num_atts': num_atts, 'bin_atts': bin_atts, 'nom_atts': nom_atts, 'ord_atts': ord_atts}
attributes = {'time_attribute': time_attribute, 'skip_attributes': skip_attributes,
'id_attribute': id_attribute, 'outcome_attribute': outcome_attribute}
return dataset, attributes, descriptives
def missing_data_method(dataset=None, descriptives=None):
print(dataset.shape)
#print(dataset.head(20))
print(dataset.isnull().sum())
# drop variables with more than 50% missing values
#perc_missing = dataset.isnull().sum() / len(dataset)
#drop_vars = perc_missing[perc_missing > 0.5].index.values
#print(drop_vars)
#print(len(drop_vars))
#smaller_dataset = dataset.drop(columns=drop_vars)
smaller_dataset = dataset.copy()
print(smaller_dataset.shape)
# drop those variables from attributes
bin_atts = descriptives['bin_atts']
nom_atts = descriptives['nom_atts']
ord_atts = descriptives['ord_atts']
for var in drop_vars:
if var in bin_atts:
bin_atts.remove(var)
elif var in nom_atts:
nom_atts.remove(var)
elif var in ord_atts:
ord_atts.remove(var)
descriptives['bin_atts'] = bin_atts
descriptives['nom_atts'] = nom_atts
descriptives['ord_atts'] = ord_atts
print(descriptives)
new_dataset = smaller_dataset.copy()
data_sorted = new_dataset.sort_values(['Wave'], ascending=[True]).reset_index(drop=True)
#print(data_sorted.tail())
return data_sorted, descriptives
def reset_attribute_type(dataset=None, descriptives=None):
finished_dataset = dataset.copy()
print(finished_dataset.dtypes)
for var in descriptives['num_atts']:
finished_dataset[var] = finished_dataset[var].astype(float)
for var in descriptives['bin_atts']:
finished_dataset[var] = finished_dataset[var].astype(object)
for var in descriptives['nom_atts']:
finished_dataset[var] = finished_dataset[var].astype(object)
# process ordinal attributes one by one
for var in descriptives['ord_atts']:
print(var)
finished_dataset[var]= finished_dataset[var].astype('category')
if var == 'Govthand':
#print(finished_dataset[var].cat.categories)
new_order = [0,3,1,2,4]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
elif var == 'Posind':
#print(finished_dataset[var].cat.categories)
new_order = [0,4,2,1,3,5]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
elif var == 'Poscountry':
#print(finished_dataset[var].cat.categories)
new_order = [0,4,2,1,3,5]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
elif var == 'profile_gross_personal':
#print(finished_dataset[var].cat.categories)
new_order = [0,10,1,3,4,5,6,7,8,9,11,12,13,2]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
elif var == 'education_age':
#print(finished_dataset[var].cat.categories)
new_order = [6,7,0,1,2,5,4,3]#[2,3,4,5,6,7,0,1]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
elif var == 'socialgradeCIE2':
#print(finished_dataset[var].cat.categories)
new_order = [0,1,2,3,4,5]
cat_type = CategoricalDtype(categories=[list(finished_dataset[var].cat.categories)[i] for i in new_order], ordered=True)
finished_dataset[var] = finished_dataset[var].astype(cat_type)
#print(finished_dataset[var].cat.categories)
print(finished_dataset.dtypes)
for var in descriptives['ord_atts']:
print(finished_dataset[var].cat.categories)
for var in descriptives['nom_atts']:
print(finished_dataset[var].unique())
for var in descriptives['bin_atts']:
print(finished_dataset[var].unique())
data_sorted = finished_dataset.sort_values(['Wave'], ascending=[True]).reset_index(drop=True)
data_sorted['id'] = np.arange(len(data_sorted))
return data_sorted