python学习（4）

05_python数据统计和可视化

01_4-1 python基本数据统计

01_1-便捷数据获取

# 用python获取数据：本地数据获取（文件的打开、读写和关闭）；网络数据获取（抓取网页，解析网页内容，urllib,urllib2,httplib,httplib2）
# 便捷网络数据
from matplotlib.finance import quotes_historical_yahoo_ochl
from datetime import date
import pandas as pd
today = date.today()
print (today)
start = (today.year-1,today.month,today.day)
# quotes = quotes_historical_yahoo_ochl ('CCE',start,today)
quotes = quotes_historical_yahoo_ochl('AXP', start, today)
df = pd.DataFrame(quotes)
print (df)

# 自然语言工具包 NLTK：古腾堡语料库，布朗语料库，路透社语料库，网络和聊天文本……
from nltk.corpus import gutenberg
import nltk
print (gutenberg.fileids())

02_2数据准备

# 数据整理
from matplotlib.finance import quotes_historical_yahoo_ochl
from datetime import date
import pandas as pd
today = date.today()
print (today)
start = (today.year-1,today.month,today.day)
# quotes = quotes_historical_yahoo_ochl ('CCE',start,today)
quotes = quotes_historical_yahoo_ochl('AMX', start, today)
fileds = ['date','open','close','high','low','volume']
df = pd.DataFrame(quotes,columns = fields)
print (df)

# 数据整理
from datetime import date
a = date.fromordinal(735190)
print (a)

2013-11-18

# 创建时间序列
import pandas as pd
dates = pd.date_range('20141001',periods=7)
print (dates)

import numpy as np
dates = pd.DataFrame(np.random.randn(7,3),index = dates,columns=list('ABC'))
print (dates)

DatetimeIndex(['2014-10-01', '2014-10-02', '2014-10-03', '2014-10-04',
               '2014-10-05', '2014-10-06', '2014-10-07'],
              dtype='datetime64[ns]', freq='D')
                   A         B         C
2014-10-01 -0.671403 -0.218888  0.238439
2014-10-02 -0.305552 -2.049908  1.416983
2014-10-03  0.438262 -0.892854  0.121548
2014-10-04 -0.966557 -0.784305 -1.017663
2014-10-05  0.882889  1.418623 -0.927695
2014-10-06 -1.236913  0.638172 -0.379498
2014-10-07  1.077244 -0.664185 -0.372371

03_3 数据显示

# 数据显示
# 显示方式：显示索引，显示列名，显示数据的值，显示数据描述
# a.index  a.columns a.values a.describe
# 索引的格式:quotesdf.index
# 显示行：专用方式；切片
# df.head[5] = df[:5]
# df.tail[5] = df[25:]

04_4数据选择

# 选择方式：选择行，选择列，选择区域，筛选（条件选择）
# 选规行：切片，索引，quotesdf[u'2013-12-02':u'2013-12-06']
# 选择列：列名，djidf['code'],djidf.code
# 选择方式：行、列 标签label（loc）  djidf.loc[1:5] djidf.loc[:['code','lasttrade']]
# 行和列的区域：标签label(loc)；单个值（at）djidf.loc[1:5，['code','lasttrade']]   djidf.loc[1，lasttrade'] djidf.at[1，lasttrade']  
# 行、列和区域：用iloc(位置)；取某个值（iat） djidf.iloc[1:6，[0,2]] djidf.iloc[1,2] djidf.iat[1,2]
# 条件筛选 quotesdf[quotesdf.index>=u'2014-01-01'] quotesdf[(quotesdf.index>=u'2014-01-01')&(quotesdf.close>=95)]

05_5简单统计与筛选

# 最近一次成交价的平均值 djidf.mean(columns = 'lasttrade')
# 最近一次成交价大于等于120的公司名 djidf[djidf.lasttrade >= 120].name
# 统计股票涨和跌的天数 len(quotesdf[quotesdf.close > quotesdf.open])  
# 统计相邻两天收盘价的涨跌情况 s=(np.sign(np.diff(quotesdf.close)) s[np.where(s == 1)].size s[np.where(s == -1)].size  
# 排序 djidf.sort(columns = 'lasttrade')  djidf.sort(columns = 'lasttrade')[27:].name
# 计数统计 统计2014年1月份的