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

13 / 32

Setting the Frequency to Business Days

It is very important for a time series to have a frequency set, so as to consolidate that the data we are working on is free of any missing data.

Since we are working with stock-exchange time-series data, it makes sense to set the frequency of the time-series to business days, meaning the data is recorded for weekdays but not any weekends.

Thanks to pandas, there is a readily available method asfreq() to set the frequency for the time-series.


asfreq() method converts the time-series to specified frequency.

We could pass an argument to the method to denote frequency. For example, df.asfreq('d') implies to set the data to a daily frequency. Similarly,

  • 'b' means business days

  • '30S' means 30 seconds

and so on.

  • First let us again check the shape of df_yahoo.

  • Now, use asfreq('b') on df_yahoo to set the frequency to business days and store it in yahoo_data data frame.

    yahoo_data = df_yahoo.<< your code comes here >>
  • Now let us see if there are any extra rows introduced after setting the frequency.

  • View the last 30 rows of the newly formed yahoo_data using tail().

    yahoo_data.<< your code comes here >>

    Observe that there are extra rows and some missing values after setting the frequency to business days.

  • Let us see the column wise null values in yahoo_data using isnull().sum().

    yahoo_data.<< you code comes here >>
Get Hint See Answer

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

Loading comments...