Get Holidays (Operator Toolbox)
SynopsisThis operator allows you to get all national holidays of a given country for a given year.
In many data science problems dates are essential. It is often very useful to know if a given date is a holiday in a given country. This additional information can be used to enhance predictive models. An example for this is given in the tutorial process.
This operator is able to return nation-wide and state-wide holidays. With default settings all holidays including state holidays are provided. To change to national only holidays you can use the 'only national holidays' parameter.
This operator is based on the 'Jollyday' library. More details can be found on https://github.com/svendiedrichsen/jollyday
- out (Data Table)
All holidays in the given years and countries as an ExampleSet.
- start_year First year (inclusive) you want to retrieve the holidays for. Range:
- end_year Last year (inclusive) you want to retrieve the holidays for. Range:
- country_name Country you want to select the holidays for. Please use 'Loop Parameters' if you want to get holidays for more than one country. Range:
- only_national_holidays If set to true, only national holidays will be returned. If set to false, the operator will also return holidays which are only a holiday in one or more state / province. Range:
- time_zone Time zone of the date. Range:
Enhance Data Set with Holidays for forecasting
In this tutorial we use 'Get Holidays' to enhance a data set with holidays for better prediction performance.
The Gas Station data set from the Samples data includes data from a single German gas station from 2017 and 2018. This tutorial process wants to forecast the next day's price. To do so we first window our data set using a 'Windowing' operator. We then join all holidays in Germany in 2017 and 2018 to the data set using nominal date representations. We generate a flag indicating any holiday and use this to enhance our forecasting using a GLM.