Question

本文档依赖的数据 option_data.csv 可以通过运行 期权高频数据准备 notebook而获取。

from matplotlib import pylab
import pandas as pd
import seaborn as sns
sns.set(style="white", context="talk")

import pandas as pd
pd.options.display.float_format = '{:,>.4f}'.format

res = pd.read_csv('option_data.csv', parse_dates=['pdDateTime'])
res['timeStamp'] = res['dataDate'] + ' ' + res['dataTime']
res['timeStamp'] = pd.to_datetime(res['timeStamp'])
res.optionId = res.optionId.astype('str')
res = res.drop('Unnamed: 0', axis=1)
res.pdDateTime = res.pdDateTime.apply(lambda x:Date(x.year,x.month,x.day))
print('开始日期： ' + res['dataDate'].iloc[0])
print('结束日期： ' + res['dataDate'].iloc[-1])
print('Market Sample: ')
res[['dataDate', 'dataTime', 'optionId', 'lastPrice', 'bidPrice1', 'askPrice1', 'lastPrice(vol)']].head()

开始日期： 2015-03-05
结束日期： 2015-03-09
Market Sample:

	dataDate	dataTime	optionId	lastPrice	bidPrice1	askPrice1	lastPrice(vol)
0	2015-03-05	09:30:00	10000001	0.1677	0.1717	0.1765	0.3468
1	2015-03-05	09:30:14	10000001	0.1717	0.1717	0.1765	0.3768
2	2015-03-05	09:30:15	10000001	0.1717	0.1610	0.1798	0.3768
3	2015-03-05	09:30:16	10000001	0.1678	0.1610	0.1798	0.3525
4	2015-03-05	09:30:18	10000001	0.1798	0.1641	0.1798	0.4205

1. 买卖价差分析

1.1 买卖价差（到期时间）

bidAskSample = res[[u'optionId', 'pdDateTime', 'dataDate', 'contractType', 'strikePrice', 'bidAskSpread(bps)']]
bidAskSample.columns = ['optionId', 'maturity', 'tradingDate', 'contractType', 'strikePrice', 'bidAskSpread(bps)']

tmp = bidAskSample.groupby(['maturity'])[['bidAskSpread(bps)']]
ax = tmp.mean().plot(kind = 'bar', figsize = (12,6), rot = 45)
ax.set_title(u'买卖价差（按照期权到期时间）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x7798290>

1.2 买卖价差（行权价）

tmp = bidAskSample.groupby(['maturity', 'strikePrice'])[['bidAskSpread(bps)']].mean().unstack()
ax = tmp.plot(kind = 'bar', figsize = (12,6), legend = True, rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = ['Strike/' + l.strip('()').split()[1] for l in labels]
ax.legend(patches, labels, loc='best', prop = font)
ax.set_title(u'买卖价差（按照期权行权价）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x5bc08d0>

1.3 买卖价差（期权类型）

tmp = bidAskSample.groupby(['maturity', 'contractType'])[['bidAskSpread(bps)']].mean().unstack()
ax = tmp.plot(kind = 'bar', figsize = (12,6), rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = [l.strip('()').split()[1] for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'买卖价差（按照期权类型）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x7a8d7d0>

2. 日交易量分析

volumeSample = res[[u'optionId', 'pdDateTime', 'dataDate', 'contractType', 'strikePrice', 'volume']]
volumeSample.columns = ['optionId', 'maturity', 'tradingDate', 'contractType', 'strikePrice', 'volume']
tmp = volumeSample.groupby(['tradingDate'])[['volume']].sum()
ax = tmp.plot(kind = 'bar', figsize = (12,6), rot = 45)
ax.set_title(u'日交易量（按交易日期）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'交易日期', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x7a72d90>

2.1 日交易量（到期时间）

tmp = volumeSample.groupby(['maturity', 'tradingDate'])[['volume']].sum().unstack()
ax = tmp.plot(kind = 'bar', figsize = (12,6), rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = [l.strip('()').split()[1] for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'日交易量（按照期权到期时间）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

每个交易日不同到期期限期权的交易量：

tmp

	volume
tradingDate	2015-03-05	2015-03-06	2015-03-09	2015-03-10	2015-03-11
maturity
March 25th, 2015	18767.0000	16704.0000	31115.0000	11888.0000	11562.0000
April 22nd, 2015	7791.0000	4468.0000	13355.0000	6909.0000	5632.0000
June 24th, 2015	965.0000	326.0000	3091.0000	619.0000	604.0000
September 23rd, 2015	635.0000	101.0000	2426.0000	240.0000	178.0000

2.2 日交易量（行权价）

tmp = volumeSample.groupby(['tradingDate','strikePrice'])[['volume']].sum().unstack()
ax = tmp.plot(kind = 'bar', figsize = (16,8),  rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = ['Strike/' + l.strip('()').split()[1] for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'日交易量（按照期权行权价）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'交易日期', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x7fa5610>

每个交易日不同行权价期权的交易量：

tmp

	volume
strikePrice	2.2000	2.2500	2.3000	2.3500	2.4000	2.4500	2.5000	2.5500
tradingDate
2015-03-05	2597.0000	1725.0000	3077.0000	5351.0000	5430.0000	4231.0000	3148.0000	2599.0000
2015-03-06	1352.0000	750.0000	1435.0000	5219.0000	4395.0000	3301.0000	3143.0000	2004.0000
2015-03-09	4576.0000	3407.0000	3599.0000	8954.0000	9564.0000	9015.0000	5969.0000	4903.0000
2015-03-10	2225.0000	1649.0000	1532.0000	3237.0000	3588.0000	2832.0000	2343.0000	2250.0000
2015-03-11	2021.0000	1286.0000	1299.0000	2959.0000	3121.0000	2648.0000	2565.0000	2077.0000

2.3 日交易量（期权类型）

tmp = volumeSample.groupby(['tradingDate','contractType'])[['volume']].sum().unstack()
ax = tmp.plot(kind = 'bar', y = ['volume'],  figsize = (12,6), rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = [l.strip('()').split()[1] for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'日交易量（按照期权类型）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'交易日期', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x8813e10>

3. 波动率价差分析

bidAskVolSample = res[[u'optionId', 'pdDateTime', 'dataDate', 'contractType', 'strikePrice', 'bidAskSpread(vol bps)']]
bidAskVolSample.columns = ['optionId', 'maturity', 'tradingDate', 'contractType', 'strikePrice', 'bidAskSpread(vol bps)']

3.1 波动率价差（到期时间）

tmp = bidAskVolSample.groupby(['maturity'])[['bidAskSpread(vol bps)']]
ax = tmp.mean().plot(kind = 'bar', figsize = (12,6), rot = 45)
ax.set_title(u'波动率价差（按照期权到期时间）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x8c0b7d0>

3.2 波动率价差（行权价）

tmp = bidAskVolSample.groupby(['maturity', 'strikePrice'])[['bidAskSpread(vol bps)']].mean().unstack()
ax = tmp.plot(kind = 'bar', figsize = (14,6), legend = True, rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = ['strike/' + l.strip('()').split()[-1] for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'波动率价差（按照期权行权价）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

3.3 波动率价差（期权类型）

tmp = bidAskVolSample.groupby(['maturity', 'contractType'])[['bidAskSpread(vol bps)']].mean().unstack()
ax = tmp.plot(kind = 'bar', figsize = (12,6), rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = [l.split()[-1].strip('()') for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'波动率价差（按照期权类型）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'到期时间', fontproperties = font, fontsize = 15)

3.4 波动率价差（交易时间）

tmp = bidAskVolSample.groupby(['tradingDate', 'maturity'])[['bidAskSpread(vol bps)']].mean().unstack()
ax = tmp.plot(kind = 'bar', figsize = (12,6), rot = 45)
patches, labels = ax.get_legend_handles_labels()
labels = [l.split(',')[1].strip('()') for l in labels]
ax.legend(patches, labels, loc='best')
ax.set_title(u'波动率价差（按照交易时间）', fontproperties = font, fontsize = 25)
ax.set_xlabel(u'交易日期', fontproperties = font, fontsize = 15)

<matplotlib.text.Text at 0x8d1fc50>

4. 个券分析

4.1 交易量

tmp = volumeSample.groupby(['tradingDate','optionId'])[['volume']].sum().unstack()
fig, axs = pylab.subplots(len(tmp)/2 + len(tmp)%2, 2, figsize = (16,8 * len(tmp)/2))
for i in range(len(tmp)):
    sample = pd.DataFrame(tmp.iloc[i]['volume'])
    sample.columns = ['volume']
    sample = sample.sort('volume', ascending = False)
    sample = sample[:10]
    row = i / 2
    col = i % 2
    sample.plot(kind = 'PIE',y = 'volume', sharex= False, ax = axs[row][col], legend = False, rot = 45)
    axs[row][col].set_title(u'交易日: ' + str(tmp.index[i]), fontproperties = font, fontsize = 18)

4.2 买卖价差

tmp = bidAskSample.groupby(['tradingDate','optionId'])[['bidAskSpread(bps)']].mean().unstack()
fig, axs = pylab.subplots(len(tmp)/2 + len(tmp)%2, 2, figsize = (16,8*len(tmp)/2))
for i in range(len(tmp)):
    sample = pd.DataFrame(tmp.iloc[i]['bidAskSpread(bps)'])
    sample.columns = ['bidAskSpread(bps)']
    sample = sample.sort('bidAskSpread(bps)')
    sample = sample[:10]
    row = i / 2
    col = i % 2
    sample.plot(kind = 'bar',y = 'bidAskSpread(bps)', sharex= False, ax = axs[row][col], legend = False, rot = 20)
    axs[row][col].set_title(u'交易日: ' + str(tmp.index[i]), fontproperties = font, fontsize = 18)

4.3 波动率价差

tmp = bidAskVolSample.groupby(['tradingDate','optionId'])[['bidAskSpread(vol bps)']].mean().unstack()
fig, axs = pylab.subplots(len(tmp)/2 + len(tmp)%2, 2, figsize = (16,8*len(tmp)/2))
for i in range(len(tmp)):
    sample = pd.DataFrame(tmp.iloc[i]['bidAskSpread(vol bps)'])
    sample.columns = ['bidAskSpread(vol bps)']
    sample = sample.sort('bidAskSpread(vol bps)')
    sample = sample[:10]
    row = i / 2
    col = i % 2
    sample.plot(kind = 'bar',y = 'bidAskSpread(vol bps)', sharex= False, ax = axs[row][col], legend = False, rot = 20)
    axs[row][col].set_title(u'交易日: ' + str(tmp.index[i]), fontproperties = font, fontsize = 18)

4.4 时间序列分析

tmp = volumeSample.groupby(['tradingDate','optionId'])[['volume']].sum().unstack()

for i, d in enumerate(tmp.index):
    fig, axs = pylab.subplots(2, 1, figsize = (16,5))
    sample = tmp.loc(d)
    sample = sample[d]
    sample.sort('volume', ascending = False)

    base = res[res['dataDate'] == d]
    base = base[base.optionId == sample.index[0][1]]
    base.index = range(len(base))

    base['calTimeStamp'] = base.timeStamp.apply(lambda s: DateTime(s.year, s.month, s.day, s.hour, s.minute, s.second))
    ax = base.plot(x = 'calTimeStamp', y = ['volume'], kind = 'bar', sharex=True, xticks = [], color = 'r', ax = axs[0])
    ax.set_title(u'交易日: ' + unicode(d) + u' 最活跃期权：'+ unicode(sample.index[0][1]), fontproperties = font, fontsize = 18)
    ax = base.plot(x= 'calTimeStamp', y = ['lastPrice(vol)'], sharex=True, legend = True,ax = axs[1], rot = 45)
    ax.set_xlabel(u'交易时间', fontproperties = font, fontsize = 15)