-
Notifications
You must be signed in to change notification settings - Fork 0
/
data.py
298 lines (236 loc) · 10 KB
/
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
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
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np
import datetime as dt
import pandas_datareader.data as web
import pickle as pkl
class Data:
'''
This class is responsible for loading the data for the VIX and Stocks
'''
def __init__(self, ticker: str, vol_file: str, read_vol_csv: bool=True, skip_init=False):
'''
:param ticker: Stock ticker to analyze
:param vol_file: corresponding volitility file
'''
self.ticker: str = ticker
self.vol_file: str = vol_file
if not skip_init:
self.end = dt.datetime.now()
# Start Year, Start Month, Start Day
self.start = dt.datetime(1990, 1, 1)
self.ticker_df = self.define_ticker_df(ticker)
self.ticker_close_array = np.asarray(self.ticker_df['Adj Close'])
self.ticker_time_array = np.asarray(self.ticker_df['Adj Close'].index)
t_arr = []
for i in self.ticker_time_array:
t_arr.append(pd.to_datetime(i, format='%Y-%m-%dT'))
self.ticker_time_array = t_arr
if read_vol_csv:
self.vix_df = self.read_vol_csv(vol_file)
close_name = ''
for col in self.vix_df.columns:
if col != 'DATE':
close_name = col
print(close_name)
self.vix_close_array = np.asarray(self.vix_df[close_name])
self.vix_time_array = np.asarray(self.vix_df['DATE'])
t_arr = []
for i in self.vix_time_array:
t_arr.append(pd.to_datetime(i,format='%Y-%m-%d'))
self.vix_time_array = t_arr
self.vix_close_array, self.vix_time_array = self.clean_vol_data()
else:
# VXX, VXZ, VIXM
self.vix_df = self.define_ticker_df(vol_file)
self.vix_close_array = np.asarray(self.vix_df['Adj Close'])
self.vix_time_array = np.asarray(self.vix_df['Adj Close'].index)
# self.plot_ticker()
print('Loaded Stock Prices:', ticker, vol_file)
self.check_length()
print("Aligned Data")
self.examples_list = self.make_data_set()
print('Created Example Set')
self.training_data, self.testing_data = self.get_test_train_split()
self.training_vix_data_split, \
self.training_spy_data_split, \
self.testing_vix_data_split,\
self.testing_spy_data_split = self.partition_data()
def partition_data(self):
'''
Creates an array of testing and training data for easy plotting
In other words, we split the data into overlapping examples, but never have all the data in one array
:return:
'''
training_vix_data_split = self.training_data[0][1][:]
training_spy_data_split = self.training_data[0][0][:]
for i in self.training_data[1:]:
training_vix_data_split.append(i[1][-1])
training_spy_data_split.append(i[0][-1])
testing_vix_data_split = self.testing_data[0][1][:]
testing_spy_data_split = self.testing_data[0][0][:]
for i in self.testing_data[1:]:
testing_vix_data_split.append(i[1][-1])
testing_spy_data_split.append(i[0][-1])
return training_vix_data_split, training_spy_data_split, testing_vix_data_split, testing_spy_data_split
def clean_vol_data(self):
'''
The CSV files return strings, and are missing a lot of values.
Replace missing with linear interpolations and convert to double
:return:
'''
vals_list = []
time_data = []
for ind, vix_val in enumerate(self.vix_close_array):
if vix_val != '.':
converted_val = np.float(vix_val)
vals_list.append(converted_val)
time_data.append(self.vix_time_array[ind])
return np.asarray(vals_list), np.asarray(time_data)
def check_length(self):
'''
Since the volitility and price data is coming from two different places, check to make sure
the same time is being compared
:return:
'''
ticker_len = len(self.ticker_close_array)
vix_len = len(self.vix_close_array)
if ticker_len > vix_len:
start_ind = self.get_start_index(self.ticker_time_array, self.vix_time_array)
self.ticker_time_array, \
self.vix_time_array, \
self.ticker_close_array,\
self.vix_close_array = self.align_timeseries(self.ticker_time_array,
self.vix_time_array,
self.ticker_close_array,
self.vix_close_array,
start_ind)
else:
start_ind = self.get_start_index(self.vix_time_array, self.ticker_time_array)
self.vix_time_array, \
self.ticker_time_array, \
self.vix_close_array, \
self.ticker_close_array = self.align_timeseries(self.vix_time_array,
self.ticker_time_array,
self.vix_close_array,
self.ticker_close_array,
start_ind)
@staticmethod
def get_start_index(longer_time_array, shorter_time_array):
'''
Finds the first data point for which the two datasets are on the same day
:param longer_time_array:
:param shorter_time_array:
:return:
'''
start_index = -1
for ind, time in enumerate(longer_time_array):
if time == shorter_time_array[0]:
print(time, shorter_time_array[0], ind)
start_index = ind
break
return start_index
@staticmethod
def align_timeseries(longer_time_array, shorter_time_array, longer_vals_arr, shorter_vals_arr, start_index):
'''
Since the data is coming from different sources, values may be missing
To avoid comparing different dates, we use a bit of dynamic time warping to make sure
that we are only including data points if they are in both data sets
'''
longer_time = []
longer_vals = []
shorter_time = []
shorter_vals = []
ind_diff = 0
for ind, time in enumerate(longer_time_array[start_index:]):
if len(shorter_time_array) > ind + ind_diff:
# One of the dates is missing so we need to shift indexes
if time != shorter_time_array[ind + ind_diff]:
print(time, shorter_time_array[ind + ind_diff])
if time > shorter_time_array[ind + ind_diff]:
ind_diff += 1
else:
ind_diff -= 1
else:
shorter_time.append(shorter_time_array[ind + ind_diff])
longer_time.append(time)
longer_vals.append(longer_vals_arr[ind + start_index])
shorter_vals.append(shorter_vals_arr[ind + ind_diff])
return longer_time, shorter_time, longer_vals, shorter_vals
@staticmethod
def replace_missing_val(arr, ind):
'''
The vol data is missing data marked by '.'
Replace with linear interpolation
'''
val = '.'
while val == '.':
val = arr[ind - 1]
ind = ind - 1
lower_val = np.float(val)
val = '.'
while val == '.':
val = arr[ind + 1]
ind = ind + 1
upper_val = np.float(val)
return (lower_val + upper_val)/2
def define_ticker_df(self, ticker):
'''
Grabs data for a given ticker from the IEX API
:param ticker:
:return:
'''
ticker_df = web.DataReader(ticker, 'yahoo', self.start, self.end)
return ticker_df
def read_vol_csv(self, filename):
df = pd.read_csv(filename)
return df
def plot_ticker(self):
df_to_plot = self.ticker_df['Adj Close']
plt.plot(df_to_plot.index, df_to_plot.values)
plt.savefig("spy.png")
plt.figure()
plt.plot(self.vix_time_array, self.vix_close_array)
plt.savefig("vix.png")
def make_data_set(self):
'''
Creates a list of tuples, where tuple[0] is the data point and tuple[1] is the target
:return:
'''
examples_list = []
for i in range(len(self.ticker_close_array[:-10])):
ticker_data = self.ticker_close_array[i:i+10]
vix_data = self.vix_close_array[i:i+10]
examples_list.append((ticker_data, vix_data))
return examples_list
def get_test_train_split(self, n=5):
'''
Splits data into fifths and creates a testing and training set.
Since we have time series data (correlated!) we cannot shuffle the examples
:param n:
:return:
'''
data_points = len(self.examples_list)
fold_length = int(data_points/n)
testing_data = self.examples_list[-fold_length:]
training_data = self.examples_list[0:(n-1) * fold_length]
return training_data, testing_data
def save_data(self):
'''
Saves Data object as pickle
'''
ticker = self.ticker
filename = 'vol_prediction' + '/' + ticker
with open(filename, 'wb') as f:
pkl.dump(self.__dict__, f)
print('Saved:', filename)
def load_data(self):
'''
Loads Data object from pickle
'''
ticker = self.ticker
filename = 'vol_prediction' + '/' + ticker
with open(filename, 'rb') as f:
temp_dict = pkl.load(f)
self.__dict__.update(temp_dict)
print('Loaded:', filename)