CODEWITHSUNDEEP
pythonpandasdataframe
AM EXPO
AM EXPO

Reputation: 75

forward min and max in time series

How to get the min and max of a time series data?
I know that the rolling gets the values from the past, however, I want to get the data for the next 3 days.

pip install yfinance

# data 
df=yf.download('tsla',start='2015-11-26',end='2021-4-28',interval='1d')

# df['min']=df['Close'].roling(3).min # this gives me the min of the past values is there a way of getting the min from the next three days?

using the shift -3 method gives me the following which does not work shift -3

Upvotes: 1

Views: 495

Answers (1)

SeaBean
SeaBean

Reputation: 23227

You can use .rolling() with parameter center=True, as follows:

df['min'] = df['Close'].rolling(7, center=True).apply(lambda x: x[-3:].min())

With parameter center=True for .rolling(), we are looking at the cener (middle) of the window. Now with window size set at 7 and looking at the middle of the window, we are working at both a backward window of size 3 and a forward window of size 3. Hence, we can use x[-3:].min() to get the min of the last 3 entries (forward 3 entries) of window x.

Result:

print(df.head(10))

                 Open       High        Low      Close  Adj Close    Volume        min
Date                                                                                  
2015-11-25  44.268002  46.166000  44.076000  45.928001  45.928001  19954000        NaN
2015-11-27  46.212002  46.450001  45.402000  46.321999  46.321999   9747000        NaN
2015-11-30  46.358002  46.855999  45.816002  46.051998  46.051998  13299000        NaN
2015-12-01  46.212002  47.599998  46.209999  47.438000  47.438000  18670000  46.076000
2015-12-02  47.400002  47.720001  46.245998  46.397999  46.397999  14907500  46.076000
2015-12-03  47.096001  47.490002  46.000000  46.542000  46.542000  14698000  45.344002
2015-12-04  46.492001  46.653999  45.532001  46.076000  46.076000  12868000  44.903999
2015-12-07  45.540001  47.125999  45.230000  46.226002  46.226002  15721000  44.903999
2015-12-08  45.504002  45.759998  44.840000  45.344002  45.344002  13438000  43.403999
2015-12-09  45.340000  45.500000  44.144001  44.903999  44.903999  15289000  43.403999

Edit

If you want to get the min of closing prices of current date plus 2 days ahead and without NaN values at the first few days, you can use numpy.fmin() as follows:

import numpy as np

df['min'] = np.fmin(np.fmin(df['Close'].values, df['Close'].shift(-1).values), df['Close'].shift(-2).values)

Result:

print(df.head(10))

                 Open       High        Low      Close  Adj Close    Volume        min
Date                                                                                  
2015-11-25  44.268002  46.166000  44.076000  45.928001  45.928001  19954000  45.928001
2015-11-27  46.212002  46.450001  45.402000  46.321999  46.321999   9747000  46.051998
2015-11-30  46.358002  46.855999  45.816002  46.051998  46.051998  13299000  46.051998
2015-12-01  46.212002  47.599998  46.209999  47.438000  47.438000  18670000  46.397999
2015-12-02  47.400002  47.720001  46.245998  46.397999  46.397999  14907500  46.076000
2015-12-03  47.096001  47.490002  46.000000  46.542000  46.542000  14698000  46.076000
2015-12-04  46.492001  46.653999  45.532001  46.076000  46.076000  12868000  45.344002
2015-12-07  45.540001  47.125999  45.230000  46.226002  46.226002  15721000  44.903999
2015-12-08  45.504002  45.759998  44.840000  45.344002  45.344002  13438000  44.903999
2015-12-09  45.340000  45.500000  44.144001  44.903999  44.903999  15289000  43.403999

Upvotes: 1

Related Questions