Data Science at Statistics Canada: Successes and Outstanding Challenges
Wesley Yung* 1
Abstract
The use of Data Science in Official Statistics has been a hot topic for several years and National Statistics Offices (NSOs) have had varying levels of success in implementing it into their statistical programs. Statistics Canada has been one of these NSOs and has successfully implemented data science methods in applications ranging from ingestion of unstructured data, coding to standard classifications, outlier detection and imputation. While there have been many successes, there are still more that can be achieved with these methods if some foundational challenges can be overcome. In the presentation, a contextualization of data science will be given, and some examples of data science methods implemented at Statistics Canada will be shared. As well, and perhaps more importantly, based on this contextualization, challenges faced and still outstanding will be presented. These challenges will cover organizational, operational, and foundational challenges.