When tracking critical changes in household income, agricultural production, school enrollment, or food security, knowing exactly when an event occurred is essential. This makes dates a crucial component of survey data. But what happens when the calendar system of your data does not align with the one most statistical software is designed to use? While the Gregorian calendar is used globally, many countries maintain their own calendar systems. Ethiopia is one of them, posing a regular data challenge to our team in Addis Ababa.
Getting to Know the Ethiopian Calendar
The Ethiopian calendar is not just a cultural artifact, it is the official system used across the country every day. The Ethiopian calendar has a unique structure making date conversion a challenging task.
- 13 Months: It consists of 12 months of 30 days, plus a 13th, short month called Pagume (5 or 6 days).
- Different Epoch: It is approximately 7–8 years behind the Gregorian calendar due to different calculations for the date of the annunciation of Jesus Christ’s birth.
- New Year: The Ethiopian New Year, Enkutatash, falls on September 11th (or 12th in a Gregorian leap year).
- Leap Year Rule: Every fourth year is a leap year without the centennial exceptions used in the Gregorian system.
The Research Problem: Manual Conversion and Analytical Risks
Since our survey respondents in Ethiopia are most familiar with the local calendar system, we frequently collect data using Ethiopian dates. While this honors local context, it creates a significant technical roadblock for analysis.
While this might seem minor, anyone who has worked with large datasets knows that problems with date formats can lead to inconsistencies and even analytical errors. This issue is particularly challenging when integrating multiple data sources, performing time-series analyses using the Ethiopian calendar, or using dates in task automation.
Common survey platforms (like SurveyCTO, KoboToolbox, and ODK) capture Ethiopian dates as simple strings or integers. Because there was no reliable, built-in solution for handling Ethiopian dates in major statistical software, converting these dates previously required analyst to manually convert thousands of entries, a time-consuming and error-prone process.
Introducing ethiodate
To solve this recurrent data challenge, Gutama, a Research Analyst at Laterite’s Ethiopia office, decided to take matters into his own hands. Gutama was inspired by the struggle he encountered firsthand: accurately computing the time gap between crop planting and harvest during his own thesis research. Recognizing the need for an efficient, standardized solution, he developed the ethiodate package for R and Stata, the two most widely used statistical software tools in social science research.
Core Functionality and Impact
- Ethiodate for R: The Ethiopian calendar R package introduces a custom ethdate class. This class works much like R’s native date object, making Ethiopian dates just as easy to manipulate and analyze as any Gregorian date. This includes adding or subtracting days, extracting components (year, month), and formatting Ethiopian dates for reports. It also includes seamless integration with ggplot2 for visualizing time-series data (see the example here). Ethiopian dates can of course also be easily converted to Gregorian dates.
- Ethiodate for Stata: The Stata Ethiopian date conversion package provides simple, precise conversion commands: to_gregorian and to_ethiopian. This streamlines the process for Stata analysts, ensuring accurate and efficient calendar conversion.
The ethiodate package provides Laterite’s research team and the wider research community with a robust toolkit to handle calendar differences effortlessly, replacing tedious manual steps with automated functions. By incorporating ethiodate into our workflows, Laterite is able to:
- Save time on data cleaning Ethiopian dates and Ethiopian to Gregorian date conversion.
- Reduce the risk of date-related errors in analysis and reporting.
- Empower local researchers who think in Ethiopian time.
Free, Open Source, Collaborative and Locally Grounded
You can find the ethiodate package for R and Stata on GitHub. It is open-source and free to use, providing a solution not only for our teams but also government agencies, NGOs, and academics worldwide who are using Ethiopian data. We welcome contributions, feedback, and bug reports as we strive to continuously improve our data analysis tools. Ultimately, the ethiodate package is more than just a conversion utility, it’s a crucial step toward making data analysis tools more inclusive and locally grounded.
If you’ve ever struggled with handling Ethiopian dates in R or Stata, ethiodate is designed to save you time, reduce errors, and instantly improve your workflow. We encourage you to try it out, check the code on GitHub, and join us in building tools that truly reflect the way we work on the ground.
This blog is written by Judith Bayer, Research Manager at Laterite.