Bolinger band ploting issue

AI & Machine learning
1

import pandas as pd

import matplotlib.pyplot as plt



def Bollinger_Bands (data, n): 

    

    MA = data['Close'].rolling(window=n).mean() 

    

    SD = data['Close'].rolling(window=n).std() 

  

    data['Lower_BB'] = MA - (2 * SD) # Lower Bollinger Band

    data['Upper_BB'] = MA + (2 * SD) # Upper Bollinger Band

   

    return data



nifty = pd.read_csv("C:/Users/jnnim/Downloads/TATACHEMBO.csv")

nifty.head()







n = 3   # IN COURSE IT IS n = 21 but when i used my data i had to maek it n = 3 why is it and what is n                         #doing in code  



nifty_bb = Bollinger_Bands(nifty, n)



nifty_bb.tail()



plt.figure(figsize=(10,5))



plt.plot(nifty_bb.Close)

plt.plot(nifty_bb.Lower_BB)

plt.plot(nifty_bb.Upper_BB)

plt.grid(True)



plt.show()

2

Here n is the size of the moving window. This is the number of observations required for calculating the statistics.



For example, if you want to calculate the rolling mean for window=3 or n=3 of a closing price.

Date
2019-12-31    157.699997
2020-01-02    160.619995
2020-01-03    158.619995
2020-01-06    159.029999
2020-01-07    157.580002
Name: Close, dtype: float64

Then the rolling function for window=3 will calculate the mean of the first three prices that is from 2019-12-31 to 2020-01-03. Then for the next period, it will calculate the mean from 2020-01-02 to 2020-01-06.

 

Date Price         Rolling Mean
2019-12-31 157.699997 NaN
2020-01-02 160.619995 NaN
2020-01-03 158.619995 158.979996
2020-01-06 159.029999 159.423330
2020-01-07 157.580002 158.409999

Here, you get two NaN values as you are calculating the rolling mean from the third period.

https://pandas.pydata.org/pandas-docs/stable/reference/api/pandas.DataFrame.rolling.html