Statistics Canada developed the Annual Survey Manufacturing and Logging Industries (ASML) Comments Classification system to automatically organize and triage respondent feedback from business surveys. The system uses advanced language processing to group comments by topic and actionability, helping analysts quickly identify and address the most important feedback.
This system is now retired and is no longer in active use. When operational, it was used by Government of Canada employees to streamline the manual review process and improve the survey experience for respondents. The system does not process personal information, focusing instead on the content and themes of comments submitted through the ASML and CAPEX surveys.
Key facts: The system uses large language model technology for zero-shot learning, meaning it could classify comments without requiring extensive training data for each category. It was designed to speed up analyst work and help identify patterns in respondent feedback for continuous improvement of Statistics Canada's survey programs.
