[FIXED] Percent change projected on a single columns

Issue

I'm trying to figure out how to project a percent variation change on a given starting amount.

To be clear, I have a dataframe which contains for each date the daily percent change. I would like to apply this percent change on a starting value in this case 100 to see its evolution over time. I have already customized the daily variation to ease the calculation as follow: Customized variation = 1 + Daily Variation/100

Amount at 100 (first row last columns is given) and this is the columns I can't find the formula to get the result

Expected Result :

I have tried many way with Pandas but it looks like it doesn't want to execute the shift on Amount columns.

I pass to my function a Pandas Serie containing the date and the daily variation in percent. Then I calculate the daily Var customised as follow: df_col_actif_var['daily_var_custo'] = (1 + df_col_actif_var/100)

And finally I try to build up the last column containing the expected results: df_col_actif_var['std'] = df_col_actif_var['std'].shift(1) * (df_col_actif_var['daily_var_custo'])

But doesn't works

Please Help !

Solution

IIUC, you can use:

df['Amount'] = (df['Daily Variation']
                    .shift(-1).cumsum()
                    .add(100).shift(1, fill_value=100))
print(df)

# Output
    Date  Daily Variation  Customized Variation  Amount
0  Day 1            -0.75                0.9925  100.00
1  Day 2             1.02                1.1020  101.02
2  Day 3            -0.19                0.9981  100.83
3  Day 4            -0.42                0.9958  100.41
4  Day 5            -0.53                0.9947   99.88

Step by step:

# shift to remove -0.75
>>> out = df['Daily Variation'].shift(-1)
0    1.02
1   -0.19
2   -0.42
3   -0.53
4     NaN

# compute cumulative sum
>>> out = out.cumsum()
0    1.02
1    0.83
2    0.41
3   -0.12
4     NaN

# add your base value 100
>>> out = out.add(100)
0    101.02
1    100.83
2    100.41
3     99.88
4       NaN

# shift back to align values and fill the first row to your base value
>>> out = out.shift(1, fill_value=100)
0    100.00
1    101.02
2    100.83
3    100.41
4     99.88

Another way to do it but I prefer to chain methods :-). However the math is much more simpler.

df['Amount'] = df.loc[1:, 'Daily Variation']
df.loc[0, 'Amount'] = 100
df['Amount'] = df['Amount'].cumsum()

Answered By - Corralien