-
Notifications
You must be signed in to change notification settings - Fork 0
/
counterfactual_no_HSR.py
252 lines (191 loc) · 11 KB
/
counterfactual_no_HSR.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
# -*- coding: utf-8 -*-
"""
Created on Tue May 22 08:50:54 2018
@author: Shih-Yang Lin
This script calculates the counterfactual market share when High Speed Rail
(HSR) does not exist.
"""
import numpy as np
import pandas as pd
from BLP_market_share import calculate_n_consumer_market_share as cal_mkt_shr
import matplotlib.pyplot as plt
import sys
# Read data
# =============================================================================
if sys.platform == 'darwin':
file_path = "/Users/errard/Dropbox/SSQ_Election/RawData20180515.csv"
market_population_file_name = "/Users/errard/Dropbox/SSQ_Election/adj_census_pop_T.csv"
xi_file_name = "/Users/errard/Dropbox/SSQ_Election/xi.csv"
else:
file_path = "C:/SSQ_Election/RawData20180515.csv"
market_population_file_name = "C:/SSQ_Election/adj_census_pop_T.csv"
xi_file_name = "C:/SSQ_Election/xi.csv"
def read_csv(file_path):
'''
This function read csv files to pandas dataframe.
input:
file_path: str.
output:
dta: pandas dataframe
'''
f = open(file_path)
dta = pd.read_csv(f)
f.close()
return(dta)
dta = read_csv(file_path)
dta['const'] = np.ones((len(dta), 1))
xi = np.loadtxt(xi_file_name)
dta = dta.assign(xi = xi)
market_population = read_csv(market_population_file_name)
market_population = market_population.drop(columns = ['depcity', 'arrcity'])
dta = pd.merge(dta, market_population,
left_on = ['depcity', 'arrcity'],
right_on = ['depcity_code', 'arrcity_code'],
how = 'left')
# Set up hyper parameters
# =============================================================================
n_consumers = int(1e4)
estimates = [-5.246357, -.6285018, -.8258803, -1.604927, 1.448219,
0.7957358, 1.093028, -0.2235259, -0.1370856, -6.126835, 1]
sigma = [2.816481]
data = dta[['priced', 'timed', 'pop_dep', 'moninc', 'edu',
'unemploymentrate_arr', 'unemploymentrate_dep',
'poll', 'mayor_arr', 'const', 'xi']]
mkt = dta['mkt']
end_var = [0]
seed = 654781324
# =============================================================================
#%% Calculate the market share when nothing changes. This calculation costs
# about 3.5 mins.
data = dta[['priced', 'timed', 'pop_dep', 'moninc', 'edu',
'unemploymentrate_arr', 'unemploymentrate_dep',
'poll', 'mayor_arr', 'const', 'xi']]
work = 'Calculating the market share when nothing changes.'
congrats = 'Finish calculating.'
dta = dta.assign(simulated_mkt_shr = cal_mkt_shr(estimates, sigma, data, end_var,
mkt, n_consumers, work, congrats, seed))
#%% Calculate counterfactual market share when HSR does not exist. This calculation
# costs about 3.5 mins.
counterfactual_data = dta[['priced', 'timed', 'pop_dep', 'moninc', 'edu',
'unemploymentrate_arr', 'unemploymentrate_dep',
'poll', 'mayor_arr', 'const', 'xi', 'brand', 'mkt']]
counterfactual_data = counterfactual_data[counterfactual_data['brand'] > 203]
mkt = counterfactual_data['mkt']
data = counterfactual_data.drop(columns = ['brand', 'mkt'])
work = 'Calculating the market share when HSR does not exist.'
counterfactual_data = counterfactual_data.assign(c_mkt_shr =
cal_mkt_shr(estimates, sigma,
data, end_var, mkt,
n_consumers, work,
congrats, seed))
del work, congrats, data, mkt
#%% Analyze the result
# remove irrelevant independent variables from counterfactual_data
counterfactual_data = counterfactual_data[['mkt', 'brand', 'c_mkt_shr']]
# add the counterfactual result in dta
dta = pd.merge(dta, counterfactual_data, left_on = ['mkt', 'brand'],
right_on = ['mkt', 'brand'], how = 'left')
# Since we do not have the counterfactual market share for HSR, the
# counterfactual market shares for HSR are missing values.
# This line fills these missing values with 0s.
dta = dta.fillna(value = {'c_mkt_shr': 0})
# Find the markets where HSR exists.
HSR_mkt = dta.loc[dta['brand'] <= 203, 'mkt'].unique()
# HSR_mkt_data is the dataset that countains the markes where HSR exists.
select = list()
for i in range(len(dta)):
select.append(dta.loc[i, 'mkt'] in HSR_mkt)
HSR_mkt_data = dta[select]
# Calculate the number of observed_voters, the number of simulated voters,
# and the number of counterfactual voters.
HSR_mkt_data = HSR_mkt_data.assign(observed_voters = HSR_mkt_data['marketshare1'] * HSR_mkt_data['census_pop'],
simulated_voters = HSR_mkt_data['simulated_mkt_shr'] * HSR_mkt_data['census_pop'],
c_voters = HSR_mkt_data['c_mkt_shr'] * HSR_mkt_data['census_pop'])
# Calculate counterfactual voters difference = the number of counterfactual
# voters - the number of simulated voters.
HSR_mkt_data['c_voters_diff'] = HSR_mkt_data['c_voters'] - HSR_mkt_data['simulated_voters']
# Create a variable, big brand:
# if brand = 201, 202, 203, b_brand = 2
# if brand = 301, 302, 303, b_brand = 3
# if brand = 401, b_brand = 4
# if brand = 502, b_brand = 5
HSR_mkt_data['b_brand'] = 0
HSR_mkt_data.loc[HSR_mkt_data['brand'] <= 203, 'b_brand'] = 2
HSR_mkt_data.loc[(HSR_mkt_data['brand'] > 203) & (HSR_mkt_data['brand'] <= 303), 'b_brand'] = 3
HSR_mkt_data.loc[HSR_mkt_data['brand'] == 401, 'b_brand'] = 4
HSR_mkt_data.loc[HSR_mkt_data['brand'] == 502, 'b_brand'] = 5
# Create a variable that indicates the available transportation modes in each
# markets.
HSR_mkt_data['available_modes'] = 0
for i in HSR_mkt_data.index:
HSR_mkt_data.loc[i, 'available_modes'] = len(HSR_mkt_data[HSR_mkt_data['mkt'] == HSR_mkt_data.loc[i, 'mkt']])
#%%
analysis_1 = HSR_mkt_data.groupby(by = ['mkt', 'b_brand']).agg({'observed_voters': 'sum',
'c_voters_diff': 'sum',
'available_modes': 'mean'})
analysis_1 = analysis_1.reset_index()
substitution_share = list()
observed_HSR_voter = list()
for i in range(len(analysis_1)):
numerator = analysis_1.loc[i, 'c_voters_diff']
denumerator = abs(analysis_1.loc[
(analysis_1['mkt'] == analysis_1.loc[i, 'mkt']) &
(analysis_1['b_brand'] == 2), 'c_voters_diff'].values)
substitution_share.append(numerator/denumerator[0])
observed_HSR_voter.append(analysis_1.loc[
(analysis_1['mkt'] == analysis_1.loc[i, 'mkt']) &
(analysis_1['b_brand'] == 2), 'observed_voters'].values[0])
del numerator, denumerator
analysis_1 = analysis_1.assign(sub_shr = substitution_share,
obs_HSR_voter = observed_HSR_voter)
analysis_1 = analysis_1.assign(predicted_voter_change = analysis_1['sub_shr'] * analysis_1['obs_HSR_voter'])
#%% Produce Table xxx
HSR_observed_voters = analysis_1[analysis_1['b_brand'] == 2]
HSR_observed_voters = HSR_observed_voters.groupby(by = ['available_modes']).agg({'observed_voters': 'sum'})
summary_result = analysis_1.groupby(by = ['available_modes']).agg({'observed_voters': 'sum',
'predicted_voter_change': 'sum'})
summary_mkt_count = analysis_1.groupby(by = ['available_modes', 'mkt']).count().reset_index()
summary_mkt_count = summary_mkt_count.groupby(by = ['available_modes']).count()
summary_result = summary_result.assign(HSR_observed_voters = HSR_observed_voters['observed_voters'],
mkt_count = summary_mkt_count['mkt'])
summary_result = summary_result.assign(avg_obs_voters = round(summary_result['observed_voters']/summary_result['mkt_count']),
avg_pre_voters_diff = round(summary_result['predicted_voter_change']/summary_result['mkt_count']),
avg_HSR_obs_voters = round(summary_result['HSR_observed_voters']/summary_result['mkt_count']))
summary_result = summary_result.assign(avg_pre_voters_diff_p = round(summary_result['avg_pre_voters_diff']*100/summary_result['avg_obs_voters'], 2))
# Calculate averages
print('The average observed HSR voters:')
print(round(sum(summary_result['HSR_observed_voters'])/sum(summary_result['mkt_count'])))
print('The average observed voters:')
print(round(sum(summary_result['observed_voters'])/sum(summary_result['mkt_count'])))
print('The average predicted voter change:')
print(round(sum(summary_result['predicted_voter_change'])/sum(summary_result['mkt_count'])))
#%% Produce Table xxx
city_location_dict = {1: 'N', 2: 'N', 3: 'N', 8: 'N',
4: 'C',
5: 'S', 6: 'S', 9: 'S'}
dep_region = analysis_1['mkt'].copy()
arr_region = analysis_1['mkt'].copy()
for i in range(len(dep_region)):
dep_region[i] = str(dep_region[i])[0]
arr_region[i] = str(arr_region[i])[2]
dep_region = dep_region.replace(city_location_dict)
arr_region = arr_region.replace(city_location_dict)
analysis_2 = analysis_1.assign(dep_region = dep_region, arr_region = arr_region)
HSR_observed_voters = analysis_2[analysis_2['b_brand'] == 2]
HSR_observed_voters = HSR_observed_voters.groupby(by = ['dep_region', 'arr_region']).agg({'observed_voters': 'sum'})
summary_result_2 = analysis_2.groupby(by = ['dep_region', 'arr_region']).agg({'observed_voters': 'sum',
'predicted_voter_change': 'sum'})
summary_mkt_count_2 = analysis_2.groupby(by = ['dep_region', 'arr_region', 'mkt']).count().reset_index()
summary_mkt_count_2 = summary_mkt_count_2.groupby(by = ['dep_region', 'arr_region']).count()
#%%
summary_result_2 = summary_result_2.assign(HSR_observed_voters = HSR_observed_voters['observed_voters'],
mkt_count = summary_mkt_count_2['mkt'])
summary_result_2 = summary_result_2.assign(avg_obs_voters = round(summary_result_2['observed_voters']/summary_result_2['mkt_count']),
avg_pre_voters_diff = round(summary_result_2['predicted_voter_change']/summary_result_2['mkt_count']),
avg_HSR_obs_voters = round(summary_result_2['HSR_observed_voters']/summary_result_2['mkt_count']),
HSR_observed_voter_mkt_shr = round(summary_result_2['HSR_observed_voters']*100/summary_result_2['observed_voters'], 2))
summary_result_2 = summary_result_2.assign(avg_pre_voters_diff_p = round(summary_result_2['avg_pre_voters_diff']*100/summary_result_2['avg_obs_voters'], 2))
# Calculate averages
print(round(sum(summary_result_2['HSR_observed_voters'])/sum(summary_result_2['mkt_count'])))
print(round(sum(summary_result_2['observed_voters'])/sum(summary_result_2['mkt_count'])))
print(round(sum(summary_result_2['predicted_voter_change'])/sum(summary_result_2['mkt_count'])))