-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathread_eurobarometer_euspeed1num.py
260 lines (211 loc) · 12.4 KB
/
read_eurobarometer_euspeed1num.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
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, skip_attributes=['euspeed1', 'lrsnum'], outcome_attribute=['euspeed1num'])
dataset, descriptives = missing_data_method(dataset=dataset, descriptives=descriptives)
finished_dataset = reset_attribute_type(dataset=dataset, trend_name=trend_name, descriptives=descriptives)
return finished_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) #71
print(len(drop_vars))
smaller_dataset = dataset.drop(columns=drop_vars)
#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)
# drop incomplete cases for age and sex, do not drop other cases
new_dataset = smaller_dataset.copy()
#print(new_dataset.isnull().sum())
new_dataset = new_dataset.dropna(how='any', axis=0, subset=['sex', 'age'])
print(new_dataset.shape)
#print(new_dataset.isnull().sum())
data_sorted = new_dataset.sort_values(['eb'], ascending=[True]).reset_index(drop=True)
data_sorted['id'] = np.arange(len(data_sorted))
#print(data_sorted.tail())
return data_sorted, descriptives
def define_attributes(data=None, skip_attributes=None, outcome_attribute=None):
time_attribute = ['eb']
#outcome_attribute = ['euspeed1num']
data_sorted = data.sort_values(['eb'], ascending=[True]).reset_index(drop=True)
id_attribute = ['id']
num_atts = ['age']
bin_atts = ['benefit', 'comm', 'ecpres', 'epinfo', 'epimp2', 'epelfut', 'mepatt', 'eurogov', 'semmedia', 'trustep', 'trustec', 'trustcm', \
'trustcj', 'trusteo', 'trustecb', 'trusteca', 'trustcr', 'trustsec', 'cpcultur', 'cpcurr', 'cpdatap', 'cpdrugs', 'cpeduc', 'cpenvir', \
'cpforpol', 'cpimmigr', 'cpindust' , 'cppasyl', 'cppress', 'cpscien', 'cpsecur', 'cpthird', 'cpunemp', 'cpvatax', 'cpwelfar', 'cpworker', \
'cpworsec', 'worldwar', 'efficacy', 'party', 'unionr', 'unionhh', 'hh', 'mhw', 'sex', 'mie']
nom_atts = ['nation1', 'epimpf', 'feel', 'sochange', 'valpri1', 'valpri2', 'closepty', 'married', 'typecmty', 'religf1', 'religf2', 'tea']
ord_atts = ['unifictn', 'membrshp', 'regret', 'commf', 'ecpresf', 'epinfof', 'epimp1', 'semhope', 'ecfinfo', 'ecint4', 'ecimp', 'citizen', 'satislfe', 'econpast', 'finapast', \
'persuade', 'polint', 'newstv', 'newspap', 'newsrad', 'relimp', 'natpride', 'satisdmo', 'satisdeu', 'mvanm', 'mvanw', 'mvecol', 'mvnatur', \
'poldisc', 'conflict', 'peaceful', 'better', 'happinss', 'underst', 'semgood', 'ecint3', 'particip', 'educ', 'educrec', 'sizehh', 'children', 'childold', 'childyng', \
'churchat', 'mediause', 'oli', 'scmi', 'lrs', 'euspeed1']
skip_attributes_temp = ['year']
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 load(trend_name=None):
name_dataset = 'ZA3521_v2-0-1_' + trend_name
location = 'data_input/Eurobarometer/' + name_dataset + '.sav'
dataset = pd.read_spss(location)
return dataset
def reset_attribute_type(dataset=None, trend_name=None, descriptives=None):
finished_dataset = dataset.copy()
#print(finished_dataset.dtypes)
finished_dataset['age'] = finished_dataset['age'].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)
print(descriptives['ord_atts'])
# make sure order of time attribute is right:
if trend_name == 'euspeed1num':
var = 'eb'
print(finished_dataset[var].cat.categories)
new_order = [0,1,2,3,4,5,6,7,8,9,10,11]
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)
if trend_name == 'lrsnum':
var = 'eb'
print(finished_dataset[var].cat.categories)
new_order = [40,49,50,51,52,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,41,42,43,44,45,46,47,48]
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)
# process ordinal attributes one by one
for var in descriptives['ord_atts']:
if var == 'membrshp':
#print(finished_dataset[var].cat.categories)
new_order = [0,2,1]
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 == 'epimp1':
#print(finished_dataset[var].cat.categories)
new_order = [1,2,0,3]
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 == 'satislfe':
#print(finished_dataset[var].cat.categories)
new_order = [1,2,0,3]
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 == 'persuade':
#print(finished_dataset[var].cat.categories)
new_order = [1,3,0,2]
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 == 'newstv':
#print(finished_dataset[var].cat.categories)
new_order = [3,2,0,4,1]
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 == 'newspap':
#print(finished_dataset[var].cat.categories)
new_order = [3,2,0,4,1]
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 == 'newsrad':
#print(finished_dataset[var].cat.categories)
new_order = [3,2,0,4,1]
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 == 'satisdmo':
#print(finished_dataset[var].cat.categories)
new_order = [1,2,0,3]
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 == 'poldisc':
#print(finished_dataset[var].cat.categories)
new_order = [1,2,0]
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 == 'educ':
#print(finished_dataset[var].cat.categories)
new_order = [9,0,1,2,3,4,5,6,7,8]
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 == 'sizehh':
#print(finished_dataset[var].cat.categories)
new_order = [4,8,7,2,1,6,5,0,3]
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 == 'children':
#print(finished_dataset[var].cat.categories)
new_order = [4,5,9,8,2,1,7,6,0,3]
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 == 'mediause':
#print(finished_dataset[var].cat.categories)
new_order = [3,1,0,2]
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 == 'oli':
#print(finished_dataset[var].cat.categories)
new_order = [1,3,2,0]
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 == 'lrs':
#print(finished_dataset[var].cat.categories)
new_order = [8,0,1,2,3,4,5,6,7,9]
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 == 'euspeed1':
#print(finished_dataset[var].cat.categories)
new_order = [6,0,1,2,3,4,5]
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.dtypes)
'''
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())
'''
return finished_dataset