动量因子、行业涨跌比率和预测收益率的数据可视化

simple_barra.py

# -*- coding:utf-8 -*-
import pandas as pd
import time, datetime
from sklearn import preprocessing
import numpy as np
import matplotlib.pyplot as plt
from pyecharts import Pie


def callist(list):
    copylist = list[:]
    copylist.sort(reverse=True)
    dic = {v: i + 1 for i, v in enumerate(copylist)}
    result = [{v: dic[v]} for i, v in enumerate(list)]
    return [float(i.values()[0]) for i in result]


wp = pd.read_table(r'J:\fi-data\analyse\whole_pic\whole_pic.txt', delimiter='\t')  ########创建wp
t1 = wp.groupby(['industry']).count().iloc[:, 1]  ########全部股票行业的频数统计
print wp['return_r'].mean(), '全部股票收益率'
wp = wp[wp.h_tr < 100]

plt.scatter(wp.h_tr, wp.return_r)
wp2 = wp[wp.h_rr > 0]
print wp['return_r'].mean(), '历史收益率大于0股票收益率'
plt.show()

model_earn = wp['return_r'][wp.h_rr + wp.h_tr > 0.1].mean()  ##########动量大于0.1的因子
print model_earn, '动量因子大于0.1股票收益率'
t2 = wp2.groupby(['industry']).count().iloc[:, 1]  ########筛选动量因子的股票行业的频数统计
t = t2 / t1
industry_factor = t.sort_values(ascending=False)  ########行业所占个股涨跌比重
wp['industry_factor'] = [industry_factor[i] for i in wp.industry.values.tolist()]  #####转化为行业因子
wp['industry_num'] = [t1[i] for i in wp.industry.values.tolist()]  #####转化为行业因子
wp['momentum'] = wp.h_tr + wp.h_rr
wp = wp[['date', 'code', 'return_r', 'h_rr', 'h_tr', 'momentum', 'industry_factor', 'industry', 'industry_num']]

plt.scatter(wp[wp['return_r'] > 0].industry_factor, wp[wp['return_r'] > 0].momentum, c='r',
                s=abs(wp.return_r[wp['return_r'] > 0] - wp.return_r[wp['return_r'] > 0].mean()) / wp[
                    wp['return_r'] > 0].return_r.std() + 10, marker='>')
plt.scatter(wp[wp['return_r'] < 0].industry_factor, wp[wp['return_r'] < 0].momentum, c='g',
                s=abs(wp[wp['return_r'] < 0].return_r - wp[wp['return_r'] < 0].return_r.mean()) / wp[
                    wp['return_r'] < 0].return_r.std() + 10, marker='<')
#plt.savefig(r'C:\\Users\\Administrator\\Desktop\\stat_anlys\\past-pattern\\' + str(
    #time.strftime(r'%Y-%m-%d', time.localtime(time.time())) + r'.png'))
plt.show()

print wp

wp = wp[wp.industry_factor < 0.75][wp.industry_factor > 0.3]
wp = wp.sort_values(by='momentum', ascending=False)
print wp['return_r'][0:10].mean(), '优势板块+动量因子的前十股票收益率'

direct = r'C:\Users\Administrator\Desktop\stat_anlys\past-pattern\last2day.txt'
wp.to_csv(path_or_buf=direct, sep='\t')