This project has been forked and modified from here.

This is the best calendar_table with the most columns of interesting date dimensions you'll ever find!

• year number, month number, day number, yearmonth, yearquarter
• month name, day name, full date w ordinal suffix (e.g., "January 3rd, 2012")
• etc.

Additionally, some interesting and unique data elements include:

• holiday identification including Easter in any year (including pre-1583)
• moon phase identification
• sunrise/sunset for UTC as well as user-provided lat/lon coordinates
• length of daylight / darkness each day/evening

For more information about the date dimensions and fields available, see these supporting documents:

• Documentation: Full column list with datatypes and descriptions
• Documentation: Sample output (download and view in a CSV viewer like Excel)

• PROCESS FLOW CONTROLS
  Check today's date. If it is a Sunday or the last day of the month, then run a code block.
• SLICING DATA FOR ANALYSIS
  From the revenue dataset, show me all sales that fall between Thanksgiving and Christmas, and determine which days be best for best door- buster promotions
• FEATURES FOR MACHINE LEARNING MODELS
  Using a polynomial linear regression, predict the end-of-month sales using the month-to-date sales along with the knowledge of what percent through the month the data is, comparing to similar data from prior months.
• DATA VISUALIZATION
  In Microsoft PowerBI, or in Excel, or in Tableau, or in any tool... Use the workday of the month as the x-axis in a bar chart that shows the change in time clocked by employees on a given project in a given month
• AND MORE?
  Send me more examples to add!

• Example CSV output is provided in git repo
• When this code is run for a span of 5 years:
  • ~2,200 rows of data are created with ~110 columns (~240k cells)
  • the resulting CSV's filesize is ~1.5 MiB
  • the script takes ~28 seconds on my raspberry pi 4B (very low specs)
  • code will run much faster on a modern laptop or desktop
• No special import statements needed besides standard Python 3.7+ packages
• 32-bit OS and hardware may cause limitations in the ability to generate table data for years far in the future (2040) and throw "OverflowError: timestamp out of range for platform time_t"; no limitation has been observed on 64-bit systems.

• modularize codes with classes/functions
• use config file to set main params (start, end, holiday rules, lat/lon for localized data like moon phase, sunrise, etc.)
• add columns
  • percent through w / ym / y / etc
  • add season based on official start/end of seasons? (equinox, etc)
• fix bugs in the following:
  • night duration utc / night duration local