stata_translate

stata_translate — Translate text variables in Stata using Google Translate API

Syntax

stata_translate varlist

Description

stata_translate takes multiple string variable (varlist) from your dataset and uses Google Translate API:

• Automatically detect the language of each row.
• Translate each row's text to English.
• Save the results in two new variables:
  • *_translated for the English translation
  • *_srclang for the detected source language

The command uses Stata’s Python integration and assumes that:

• The translation service is online Google Translate API
• The model llama3.2:latest for local / offline usage will be developed later using ollama

Example

clear
input str40 city_description str40 food_review
París es una ciudad hermosa" "Le fromage est délicieux"
Berlin ist großartig" "Ich liebe Bratwurst"
東京はとても忙しいです" "寿司はとても新鮮です"
end

stata_translate city_description food_review

list city_description food_review

This will create two new variables:

• *_translated — English translation
• *_srclang — Source language (e.g., 'de', 'fr', 'id')

Requirements

To use this command, you must:

1. Have Python configured in Stata (see python query). The easiest way is to install Anaconda or Miniconda to your environment (see https://www.anaconda.com/download)
2. Install the following Python libraries: (this version of stata_translate automatically install these packages as long as you have Python environment installed)
   • langdetect (pip install langdetect)
   • deep_translator (pip install deep_translator)

Author

Written by @akirawisnu during his PhD depression break.
Feedback or improvements welcome!

Also see

• Manual: help python, help generate, help string functions
• Online: https://ollama.com, https://pypi.org/project/langdetect, https://www.anaconda.com/download