使用数据框列值的 Python Pandas 图
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/42744189/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Python Pandas plot using dataframe column values
提问by LKM
I'm trying to plot a graph using dataframes.
我正在尝试使用数据框绘制图形。
I'm using 'pandas_datareader' to get the data.
我正在使用“pandas_datareader”来获取数据。
so my code is below:
所以我的代码如下：
tickers = ["AAPL","GOOG","MSFT","XOM","BRK-A","FB","JNJ","GE","AMZN","WFC"]
import pandas_datareader.data as web
import datetime as dt
end = dt.datetime.now().strftime("%Y-%m-%d")
start = (dt.datetime.now()-dt.timedelta(days=365*3)).strftime("%Y-%m-%d")
%matplotlib inline
import matplotlib.pyplot as plt
import pandas as pd
data = []
for ticker in tickers:
sub_df = web.get_data_yahoo(ticker, start, end)
sub_df["name"] = ticker
data.append(sub_df)
data = pd.concat(data)
So in the variable data, there are 8 columns = ['Date', 'Open', 'High' ,'Low' ,'Close' 'Volume', 'Adj Close','name']
所以在变量中data，有 8 列 =['Date', 'Open', 'High' ,'Low' ,'Close' 'Volume', 'Adj Close','name']
The variable 'data' is shown below:
变量“数据”如下所示：
What I want to do is to plot a graph taking 'date' values as x-parameter , 'high' as y-parameter with multiple columns as 'name' column values(=["AAPL","GOOG","MSFT","XOM","BRK-A","FB","JNJ","GE","AMZN","WFC"]).
我想要做的是绘制一个图形，将“日期”值作为 x 参数，“高”作为 y 参数，将多列作为“名称”列值（=[“AAPL”，“GOOG”，“MSFT” 、“XOM”、“BRK-A”、“FB”、“JNJ”、“GE”、“AMZN”、“WFC”]）。
How can I do this?
我怎样才能做到这一点？
When i executed data.plot(), the result takes dataas x-parameter well but there are 5 columns ['open','high','low','close','volume','adj close']not 7 columns ["AAPL","GOOG","MSFT","XOM","BRK-A","FB","JNJ","GE","AMZN","WFC"]: what i want to do.
The result is below:
当我执行时data.plot()，结果data很好地作为 x 参数，但有 5 列['open','high','low','close','volume','adj close']而不是 7 列["AAPL","GOOG","MSFT","XOM","BRK-A","FB","JNJ","GE","AMZN","WFC"]：我想要做什么。结果如下：
回答by Psidom
You need to reshape your data so that the names become the header of the data frame, here since you want to plot Highonly, you can extract the Highand namecolumns, and transform it to wide format, then do the plot:
您需要重塑数据，使名称成为数据框的标题，因为您只想绘制High，因此可以提取High和name列，并将其转换为宽格式，然后进行绘图：
import matplotlib as mpl
mpl.rcParams['savefig.dpi'] = 120
high = data[["High", "name"]].set_index("name", append=True).High.unstack("name")
# notice here I scale down the BRK-A column so that it will be at the same scale as other columns
high['BRK-A'] = high['BRK-A']/1000
high.head()
ax = high.plot(figsize = (16, 10))
回答by MaxU
@Psidomand @Grrhave already given you very good answers.
I just wanted to add that pandas_datareaderallows us to read all data into a Pandas.Panel conviniently in one step:
我只是想补充一点，它pandas_datareader允许我们一步轻松地将所有数据读入 Pandas.Panel：
p = web.DataReader(tickers, 'yahoo', start, end)
now we can easily slice it as we wish
现在我们可以根据需要轻松切片
# i'll intentionally exclude `BRK-A` as it spoils the whole graph
p.loc['High', :, ~p.minor_axis.isin(['BRK-A'])].plot(figsize=(10,8))
alternatively you can slice on the fly and save only Highvalues:
或者，您可以即时切片并仅保存High值：
df = web.DataReader(tickers, 'yahoo', start, end).loc['High']
which gives us:
这给了我们：
In [68]: df
Out[68]:
AAPL AMZN BRK-A FB GE GOOG JNJ MSFT WFC XOM
Date
2014-03-13 539.659988 383.109985 188852.0 71.349998 26.000000 1210.502120 94.199997 38.450001 48.299999 94.570000
2014-03-14 530.890015 378.570007 186507.0 69.430000 25.379999 1190.872020 93.440002 38.139999 48.070000 94.220001
2014-03-17 529.969994 378.850006 185790.0 68.949997 25.629999 1197.072063 94.180000 38.410000 48.169998 94.529999
2014-03-18 531.969986 379.000000 185400.0 69.599998 25.730000 1211.532091 94.239998 39.900002 48.450001 95.250000
2014-03-19 536.239990 379.000000 185489.0 69.290001 25.700001 1211.992061 94.360001 39.549999 48.410000 95.300003
2014-03-20 532.669975 373.000000 186742.0 68.230003 25.370001 1209.612076 94.190002 40.650002 49.360001 94.739998
2014-03-21 533.750000 372.839996 188598.0 67.919998 25.830000 1209.632048 95.930000 40.939999 49.970001 95.989998
... ... ... ... ... ... ... ... ... ... ...
2017-03-02 140.279999 854.820007 266445.0 137.820007 30.230000 834.510010 124.360001 64.750000 59.790001 84.250000
2017-03-03 139.830002 851.989990 264690.0 137.330002 30.219999 831.359985 123.930000 64.279999 59.240002 83.599998
2017-03-06 139.770004 848.489990 263760.0 137.830002 30.080000 828.880005 124.430000 64.559998 58.880001 82.900002
2017-03-07 139.979996 848.460022 263560.0 138.369995 29.990000 833.409973 124.459999 64.779999 58.520000 83.290001
2017-03-08 139.800003 853.070007 263900.0 137.990005 29.940001 838.150024 124.680000 65.080002 59.130001 82.379997
2017-03-09 138.789993 856.400024 263620.0 138.570007 29.830000 842.000000 126.209999 65.199997 58.869999 81.720001
2017-03-10 139.360001 857.349976 263800.0 139.490005 30.430000 844.909973 126.489998 65.260002 59.180000 82.470001
[755 rows x 10 columns]
回答by Grr
You should group your data by nameand then plot. Something like data.groupby('name').plot ()should get you started. You may need to feed in dateas the x value and highfor the y. Cant test it myself at the moment as i am on mobile.
您应该将数据分组name，然后进行绘图。类似的东西data.groupby('name').plot ()应该让你开始。您可能需要date输入 x 值和highy值。由于我在移动设备上，目前无法自己测试。
Update
更新
After getting to a computer this I realized I was a bit off. You would need to reset the index before grouping then plot and finally update the legend. Like so:
拿到电脑后，我意识到我有点不对劲。您需要在分组之前重置索引，然后绘制并最终更新图例。像这样：
fig, ax = plt.subplots()
names = data.name.unique()
data.reset_index().groupby('name').plot(x='Date', y='High', ax=ax)
plt.legent(names)
plt.show()
Granted if you want this graph to make any sense you will need to do some form of adjustment for values as BRK-A is far more expensive than any of the other equities.
当然，如果您希望此图表有意义，您将需要对价值进行某种形式的调整，因为 BRK-A 比任何其他股票都贵得多。
