python-for-finances/daily-and-cumulative-return/daily-and-cumulative-return.ipynb

Libraries installation

In [18]:

!pip install pandas -q
!pip install numpy -q
!pip install yfinance -q
!pip install matplotlib -q

In [4]:

import pandas as pd
import numpy as np
import yfinance as yf
from datetime import datetime

Building investment wallet

In [13]:

# Period
start = "2018-01-01"
end = f"{datetime.now():%Y-%m-%d}"

# Selection of wallet assets
assets = ["BBAS3.SA", "B3SA3.SA", "ITSA4.SA", "PETR4.SA", "VALE3.SA",
          "TRPL4.SA", "EGIE3.SA", "TAEE11.SA", "WEGE3.SA", "KLBN4.SA"]

# Assignment of 10% weight for each asset
weights = np.array([.1] * 10)

Downloading financial data

In [15]:

wallet = yf.download(assets, start, end)["Adj Close"]

[*********************100%***********************]  10 of 10 completed

In [16]:

wallet.head()

Out [16]:

	B3SA3.SA	BBAS3.SA	EGIE3.SA	ITSA4.SA	KLBN4.SA	PETR4.SA	TAEE11.SA	TRPL4.SA	VALE3.SA	WEGE3.SA
Date
2018-01-02	6.221113	23.474606	18.884989	6.252337	2.534116	6.091883	12.486040	10.113166	28.455339	8.858922
2018-01-03	6.327015	23.774002	18.758101	6.326030	2.510867	6.147096	12.503497	9.907570	28.284828	8.724374
2018-01-04	6.367747	24.002117	18.514900	6.445067	2.541866	6.158139	12.305675	9.635947	28.400774	8.639398
2018-01-05	6.460073	24.002117	18.610065	6.456405	2.495369	6.194947	12.363855	9.790517	28.844112	8.674805
2018-01-08	6.489944	24.059149	18.715805	6.450731	2.503118	6.268565	12.305675	9.739493	29.485239	8.745622

In [23]:

# Transform date into base 1 relative scale
(wallet / wallet.iloc[0]).plot();

Daily and cumulative return

In [26]:

returns = wallet.pct_change()

In [27]:

# Weighted return, that means, how much each asset acted in the return of the wallet
returns * weights

Out [27]:

	B3SA3.SA	BBAS3.SA	EGIE3.SA	ITSA4.SA	KLBN4.SA	PETR4.SA	TAEE11.SA	TRPL4.SA	VALE3.SA	WEGE3.SA
Date
2018-01-02	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN
2018-01-03	0.001702	0.001275	-0.000672	0.001179	-0.000917	0.000906	0.000140	-0.002033	-0.000599	-0.001519
2018-01-04	0.000644	0.000960	-0.001297	0.001882	0.001235	0.000180	-0.001582	-0.002742	0.000410	-0.000974
2018-01-05	0.001450	0.000000	0.000514	0.000176	-0.001829	0.000598	0.000473	0.001604	0.001561	0.000410
2018-01-08	0.000462	0.000238	0.000568	-0.000088	0.000311	0.001188	-0.000471	-0.000521	0.002223	0.000816
...	...	...	...	...	...	...	...	...	...	...
2023-07-31	0.001223	0.001925	-0.000989	0.000102	0.001535	0.004536	-0.000525	0.000120	0.002262	-0.000992
2023-08-01	-0.001074	-0.001722	0.000613	-0.000204	-0.001944	-0.001639	0.000361	0.001394	-0.001388	-0.000150
2023-08-02	0.000204	0.000443	0.000068	0.000510	0.000881	-0.000229	0.000194	0.000747	-0.001628	0.000426
2023-08-03	0.000677	0.000168	-0.001376	-0.000102	0.001747	0.001277	0.000608	0.000078	0.000626	0.000999
2023-08-04	-0.001076	-0.000525	-0.000617	-0.000813	0.001931	-0.002975	0.000220	-0.000429	0.000074	0.001039

1388 rows × 10 columns

In [29]:

wallet_return = (returns * weights).sum(axis = 1)

In [31]:

wallet_return.plot();

In [32]:

# Cumulative product on base 0
cumulative_return_wallet = (1 + wallet_return).cumprod() - 1

In [34]:

cumulative_return_wallet.plot();

In [44]:

ibov = yf.download("^BVSP", start, end)["Adj Close"]
ibov_return = ibov.pct_change()
cumulative_return_ibov = (1 + ibov_return).cumprod() - 1

[*********************100%***********************]  1 of 1 completed

In [46]:

cumulative_return_ibov.plot();

Comparing performance of IBOV x Wallet

In [47]:

portifolio_return = pd.DataFrame()
portifolio_return["Wallet Return"] = cumulative_return_wallet
portifolio_return["IBOV Return"] = cumulative_return_ibov
portifolio_return = portifolio_return.dropna()
portifolio_return

Out [47]:

	Wallet Return	IBOV Return
Date
2018-01-03	-0.000538	0.001335
2018-01-04	-0.001822	0.009706
2018-01-05	0.003125	0.015149
2018-01-08	0.007866	0.019104
2018-01-09	0.004247	0.012492
...	...	...
2023-07-31	2.119721	0.565560
2023-08-01	2.101773	0.556637
2023-08-02	2.106787	0.551643
2023-08-03	2.121399	0.548138
2023-08-04	2.111500	0.534298

1386 rows × 2 columns

In [48]:

portifolio_return.plot();

In [ ]: