Project - Stock Closing Price Prediction using Deep Learning, TensorFlow2 & Keras

15 / 32
Previous Index Next  

Filtering out the Holidays

Let us check if the dates of missing values fall in any holidays official for NYSE. A quick check at the NYSE official website would help us know the list of official holidays of NYSE.

Now that we know the list of official holidays for NYSE, we use the calendar module and datetime module to check if a given date in the null_dates list is a holiday or not. We do this by:

  • Determining if that week or day name of the date in that month is a holiday or not. For example, Washington's Birthday is celebrated on the third Monday of February.
  • So we extract the week, day, month, year from a given date in null_dates, and check if the day falls between 15 and 21(the possible day range of third Monday) and if the week is Monday.
  • If yes, then it is a holiday. Then we append that date in the holidays list. We took a list of the dates when Good Friday occurs since there is no particular way to determine it.
  • Finally, after obtaining the entire list of holidays, we filter out the holidays dates from null_dates and store the non-holiday dates in the list non_holidays.
INSTRUCTIONS

Copy-paste the following code to get the non_holidays, which indicate the dates where the values are missing other than holidays.

import calendar
import datetime

holidays = []

for date in null_dates:
    week, day, month, year = date.weekday(), date.day, date.month, date.year
    week_day = calendar.day_name[week]

    if month==1:
        if day==1:
            # New year day
            holidays.append(date)
        elif day==2 and week_day=='Monday':
            # Observed New Year Day
            holidays.append(date)
        elif day>=15 and day<=21 and week_day=='Monday':
            # Martin Luther King, Jr. Day
            holidays.append(date)

    elif month==2:
        # Washington's Birthday
        if day>=15 and day<=21 and week_day=='Monday':
            holidays.append(date)

    elif month==5:
        # Memorial day
        if day>=25 and day<=31 and week_day=='Monday':
            holidays.append(date)

    elif month==7:
        # Independence day
        if day==4:
            holidays.append(date)
        # Observed Independence  Day
        elif day==5 and week_day=='Monday':
            holidays.append(date)
        elif day==3 and week_day=='Friday':
            holidays.append(date)

    elif month == 9:
        # Labour day
        if day>=1 and day<=7 and week_day=='Monday':
            holidays.append(date)

    elif month==11:
        # Thanksgiving Day
        if week_day=='Thursday' and day>=22 and day<=28:
            holidays.append(date)

    elif month==12:
        # Christmas Day
        if day==25:
            holidays.append(date)
        # Observed Christmas Day
        elif day==24 and week_day=='Friday':
            holidays.append(date)
        elif day==26 and week_day=='Monday':
            holidays.append(date)

good_fridays = [ datetime.date(2010,4,2), datetime.date(2011,4,22), datetime.date(2012,4,6), datetime.date(2013,3,29), datetime.date(2014,4,18), datetime.date(2015,4,3), datetime.date(2016,3,25) ]
holidays = holidays + [pd.to_datetime(date) for date in good_fridays]

non_holidays = [x for x in null_dates if x not in holidays]
print(non_holidays)

We observe that the values are missing for October 29th and 30th of 2012.

Get Hint See Answer


Note - Having trouble with the assessment engine? Follow the steps listed here

Previous Index Next  

Loading comments...