蟒蛇生长因子的计算

Question

我有以下数据

Date           Country  Sales
01/02/20        France   1000
01/02/20        Germany  1000
02/02/20        France   2000
02/02/20        Germany  3000
03/02/20        France   4500
03/02/20        Germany  5000

我想按国家计算每天的增长因子。例如，德国02/02/20 = 3000/1000和03/02/2020的增长率应为5000/3000。不是在找百分比的变化。

Answer 1

df = pd.DataFrame([
['01/02/20',        'France '  ,1000],
['01/02/20',        'Germany'  ,1000],
['02/02/20',        'France '  ,2000],
['02/02/20',        'Germany'  ,3000],
['03/02/20',        'France '  ,4500],
['03/02/20',        'Germany'  ,5000]
], columns=['Date', 'Country', 'Sales'])
df['Diff'] = df.groupby(['Country']).diff()
df['Growth'] = df.eval('Sales/(Sales - Diff)')
print(df)

另一种选择是如上所述。

       Date  Country  Sales    Diff    Growth
0  01/02/20  France    1000     NaN       NaN
1  01/02/20  Germany   1000     NaN       NaN
2  02/02/20  France    2000  1000.0  2.000000
3  02/02/20  Germany   3000  2000.0  3.000000
4  03/02/20  France    4500  2500.0  2.250000
5  03/02/20  Germany   5000  2000.0  1.666667

Answer 2

从按国家进行排序开始：df = df.sort_values(['Country'])

然后设置一个过滤器，它只在索引与前面的索引相同(即它们来自同一个国家)时返回true：filter = df['Country'] == df['Country'].shift(1)

现在，在上面的过滤器为真的最后一点上计算分数变化：

df['Growth'] = df['Sales'].where(filter)/df['Sales'].shift(1)

最后，按指数计算：

df = df.sort_index()

在以下方面的成果：

       Date  Country  Sales     Growth
0  01/02/20   France   1000         NaN
1  01/02/20  Germany   1000         NaN
2  02/02/20   France   2000    2.000000
3  02/02/20  Germany   3000    3.000000
4  03/02/20   France   4500    2.250000
5  03/02/20  Germany   5000    1.666667

Answer 3

使用：

df["Growth Factor"] = df.groupby(
    "Country")["Sales"].transform(lambda g: g / g.shift())
print(df)

这些指纹：

       Date  Country  Sales  Growth Factor
0  01/02/20   France   1000            NaN
1  01/02/20  Germany   1000            NaN
2  02/02/20   France   2000       2.000000
3  02/02/20  Germany   3000       3.000000
4  03/02/20   France   4500       2.250000
5  03/02/20  Germany   5000       1.666667