-
Notifications
You must be signed in to change notification settings - Fork 0
/
JM_datacleaning.py
231 lines (185 loc) · 8.05 KB
/
JM_datacleaning.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
import numpy as np
import pandas as pd
import os
import gc
import myfunc
import re
from matplotlib import pyplot as plt
from sklearn.linear_model import LinearRegression
from fuzzywuzzy import process
######################################导入源数据
data = pd.read_excel('e:\\jm\\附件1.xlsx')
data.to_csv('e:\\jm\\data1.csv')
data1 = pd.read_csv('e:\\jm\\data1.csv')
data2 = pd.read_sas('e:\\jm\\data.sas7bdat')
myfunc.search('map')
#######################################导入数据字典
lineData = []
with open('e:\\jm\\数据字典.txt') as txtData:
lines = txtData.readlines()
for line in lines:
lineData = lineData + [line.strip()]
def var_dic(n):
var_lx = lineData[n].replace(')', '）').replace('(', '（').replace(' ', '').split('、')
t = [re.split('[（）]', x) for x in var_lx]
var_lx = dict()
for x in t:
var_lx[x[1]] = x[0]
return var_lx
var_lx = var_dic(0) # 连续变量字典
var_str = var_dic(1) # 字符变量字典
var_ls = var_dic(2) # 离散变量字典
var_dictionary = dict(var_lx, **var_str, **var_ls)
var_dictionary['guncertain1']
#########################各个指标的缺失值比率统计
var = data.columns
var_N = len(var)
length = len(data)
def find_miss(data):
var_missRate = []
var = data.columns
for x in var:
t = sum(pd.isnull(data[x])) / length
var_missRate = var_missRate + [t]
return pd.Series(var_missRate, index=var)
var_missRate = find_miss(data)
#########################将缺失值比率大于80%的变量进行删除
print(len(var[np.where(var_missRate > 0.8)]))
var_del = var[np.where(var_missRate > 0.8)]
var = var[np.where(var_missRate <= 0.8)]
data1 = data[var]
#########################描述字符型变量的赋值
data1.eventid = data1.eventid.astype(str)
summary = data1.pop('summary')
target1 = data1.pop('target1')
data2 = data1.drop(['location', 'corp1', 'natlty1_txt', 'motive', 'weapdetail',
'propcomment', 'addnotes', 'scite1', 'scite2', 'scite3', 'dbsource'], axis=1)
# 记录字符型变量的对映值
# 国家名称：美国、英国等
country = data2[['country', 'country_txt']].sort_values('country').drop_duplicates('country')
country.to_csv('e:\\jm\\country.csv',encoding='gbk')
# 区域名称：北美、东亚等
region = data2[['region', 'region_txt']].sort_values('region').drop_duplicates('region')
print(region)
# 攻击类型：暗杀、劫持等
attacktype1 = data2[['attacktype1', 'attacktype1_txt']].sort_values('attacktype1').drop_duplicates('attacktype1')
# 目标类型：商业、政府等
targtype1 = data2[['targtype1', 'targtype1_txt']].sort_values('targtype1').drop_duplicates('targtype1')
# 目标子类型：银行、酒店等
targsubtype1 = data2[['targsubtype1', 'targsubtype1_txt']].sort_values('targsubtype1').drop_duplicates('targsubtype1')
# 武器类型：轻武器、爆炸物等
weaptype1 = data2[['weaptype1', 'weaptype1_txt']].sort_values('weaptype1').drop_duplicates('weaptype1')
# 武器子类型：炸药、车辆等
weapsubtype1 = data2[['weapsubtype1', 'weapsubtype1_txt']].sort_values('weapsubtype1').drop_duplicates('weapsubtype1')
# 财产损失：轻微损失、损失不明等
propextent = data2[['propextent', 'propextent_txt']].sort_values('propextent').drop_duplicates('propextent').dropna()
####对上述有映射的字符说明指标删除
data2 = data2.drop(['country_txt', 'region_txt', 'attacktype1_txt', 'targtype1_txt',
'targsubtype1_txt', 'weaptype1_txt', 'weapsubtype1_txt', 'propextent_txt'], axis=1)
######################进一步考察缺失值状况
var_missRate = find_miss(data2)
# 考虑到武器子类型、目标子类型并不重要且缺失值6~7%予以删除
data2 = data2.drop(['targsubtype1', 'weapsubtype1'], axis=1)
# 财产损失用0值填充
data2[['propextent', 'propvalue']] = data2[['propextent', 'propvalue']].fillna(0)
# 相关人数按字典要求用-99填充
data2[['nperps', 'nperpcap']] = data2[['nperps', 'nperpcap']].fillna(-99)
# 当前缺失值大于1%，仅伤亡人数4个指标
print(find_miss(data2)[find_miss(data2) > 0.01])
# 导出经纬度
# t = data2[['latitude', 'longitude','nkill']].dropna()
# t=t.groupby(['latitude', 'longitude']).sum()
# t = t.reset_index()
# t=t[t.nkill>0]
# t.to_csv('e:\\jm\\jwd.csv',index=False)
# 相关性，填充 受伤人数
def fill_var(data2, var1, var2):
# var1 是自变量 var2 是因变量，对缺失部分的因变量进行填充
print(data2[[var1, var2]].corr())
# 回归
t = data2[[var1, var2]].dropna()
lrModel = LinearRegression()
lrModel.fit(t[var1].values.reshape(-1, 1), t[var2])
a, b = lrModel.coef_, lrModel.intercept_
# 填充nwound缺失部分
t1 = data2[(~data2[var1].isnull()) & (data2[var2].isnull())][[var1, var2]]
data2.loc[t1.index, var2] = lrModel.predict(t1[var1].values.reshape(-1, 1))
return data2
data2 = fill_var(data2, 'nkill', 'nwound')
data2 = fill_var(data2, 'nkillus', 'nwoundus')
# 填充
data3 = data2.fillna(-1)
len(data3)
data2.to_csv('e:\\jm\\data2.csv',encoding='utf-8')
data3 = data2.dropna()
print('清洗前数据量：', len(data2))
print('清洗后数据量：', len(data3))
#######################先不删除缺失值
###将数据集中多个分类数字修正为int型
var_int = list(set(data2.columns) - {'eventid', 'provstate', 'city', 'latitude', 'longitude', 'gname'})
data3[var_int] = data3[var_int].astype(int)
# 保存数据集
data3.to_csv('e:\\jm\\data3.csv')
#########################导入各国经济自由度数据集
eco = pd.read_excel('e:\\jm\\全球各国经济自由度指数.xlsx')
eco.columns = ['国家', '年份', '总体评分'] + eco.iloc[1, 3:].values.tolist()
eco = eco[2:]
#填充
a=eco.pivot(index='年份',columns='国家')
for column in list(a.columns[a.isnull().sum() > 0]):
mean_val = a[column].mean()
a[column].fillna(mean_val, inplace=True)
eco=a.stack(dropna=False).reset_index()
help(pd.DataFrame.stack)
# 对国家进行匹配
cou1 = eco[['国家']].sort_values('国家').drop_duplicates('国家')
cou2 = country.sort_values('country_txt')
# 原数据中名称字符分隔符删除
cou2['country_txt'] = cou2['country_txt'].str.replace(' ', '')
cou = pd.merge(cou1, cou2, left_on=['国家'], right_on=['country_txt'], how='outer')
# 将不能匹配的进行模糊匹配
a = cou[cou.isnull().any(axis=1)]
a1 = a['国家'].dropna().values.tolist()
a2 = a['country_txt'].dropna().values.tolist()
dic = dict()
for t in a2:
x = process.extract(t, a1, limit=2)
if (x[0][1] >= 68 or x[0][1] == 50) and (x[0][1] != 72):
dic[t] = x[0][0]
t = [a for a, b in enumerate(cou['country_txt']) if b in dic]
cou.iloc[t, 0] = list(dic.values())
cou = cou.iloc[:,:2].dropna()
cou['country']=cou['country'].astype(int)
# 将国家编号并入eco经济数据集
eco1 = pd.merge(eco, cou, on=['国家'], how='right')
eco1.rename(columns={'年份':'iyear'},inplace=True)
eco1 = eco1.drop('国家',axis=1)
#将经济数据按照 国家编号、年份汇入 恐怖袭击 表格
eco1 = eco1.sort_values(['iyear', 'country'])
data4 = pd.merge(data3, eco1, on=['country', 'iyear'], how='left')
############################################
find_miss(data4)
data4=data4.drop(['司法效率','财政健康'],axis=1)
print(len(data4.dropna()))
len(data4)
data4.to_csv('e:\\jm\\data4.csv')
data4.to_csv('e:\\jm\\data4_1.csv')
####################第二问：对未知的分类
g = pd.read_excel('e:\\jm\\nogroup(1).xlsx',head=False)
g.eventid = g.eventid.astype('str')
data4 = pd.merge(g,data4,on='eventid',how='inner')
len(g)
np.shape(data3)
t = pd.Series(var_dictionary)
t911 = data2[data2.eventid.str.startswith('20010911')]
t911 = data3[data3.eventid.str.startswith('20010911')]
t = t[set(t.index) & set(a)]
del con2
t1 = data1[data1.country_txt == 'China']
data1.discribe()
myfunc.search('rename')
help(pd.merge)
data2.to_excel('e:\\jm\\data2.xlsx', 'sheet1')
data2.eventid[1]
myfunc.search('to_excel')
data.shape