'Inverse' cumprod in pandas

Question

I have a data frame which contains dates as index and a value column storing growth percentage between consecutive dates (i.e. dates in the index). Suppose I want to compute 'real' values by setting a 100 basis at the first date of the index and then iteratively applying the % of growth. It is easy with the cumprod method.

Now, I want to set as 100 basis the laste date in the index. I thus need to compute for each date in the index the 'inverse' growth. Is there an easy way (and pythonic) to do this with pandas?

Regards,

Allia

Answer 1

Consider the dataframe df with random returns for 10 days.

np.random.seed([3,1415])
tidx = pd.date_range('2012-04-01', periods=10)
df = pd.DataFrame(dict(A=np.random.rand(10) / 10), tidx)

df

                   A
2012-04-01  0.044494
2012-04-02  0.040755
2012-04-03  0.046015
2012-04-04  0.046524
2012-04-05  0.046269
2012-04-06  0.001655
2012-04-07  0.085045
2012-04-08  0.081774
2012-04-09  0.077796
2012-04-10  0.075798

---

You want to use cumprod then divide by the last value

dcum = df.add(1).cumprod()
dcum.div(dcum.iloc[-1]).mul(100)

                     A
2012-04-01   61.539104
2012-04-02   64.047157
2012-04-03   66.994277
2012-04-04   70.111111
2012-04-05   73.355090
2012-04-06   73.476459
2012-04-07   79.725230
2012-04-08   86.244715
2012-04-09   92.954225
2012-04-10  100.000000