date or other data normalization general solution in pandas

import pandas as pd

import numpy as np

from sklearn.preprocessing import MinMaxScaler

import time

 

# of course you can use basic pandas api doing this job, but I'd prefer general solution hereby

def convert_to_timestamp(x):

    """Convert date objects to integers"""

    return time.mktime(x.to_datetime().timetuple())

def normalize(df):

    """Normalize the DF using min/max"""

    scaler = MinMaxScaler(feature_range=(-1, 1))

    dates_scaled = scaler.fit_transform(df['dates'])

    return dates_scaled

 

 

if __name__ == '__main__':

    # Create a random series of dates

    df = pd.DataFrame({

        'dates':

            ['1980-01-01', '1980-02-02', '1980-03-02', '1980-01-21',

             '1981-01-21', '1991-02-21', '1991-03-23']

    })

    # Convert to date objects

    df['dates'] = pd.to_datetime(df['dates'])

    # Now df has date objects like you would, we convert to UNIX timestamps

    df['dates'] = df['dates'].apply(convert_to_timestamp)

    # Call normalization function

    df = normalize(df)