Jakarta, Indonesia – The “Machine Learning for Official Statistics” Workshop, held from February 3–7, 2025, in Jakarta, was successfully conducted, bringing together statisticians and data scientists from across the Asia-Pacific region. The event was organized by Statistics Indonesia (BPS), the United Nations Statistics Division (UNSD), and the Statistical Institute for Asia and the Pacific (SIAP), aiming to strengthen statistical capacity through machine learning techniques.

A Step Towards More Modern Official Statistics

With the growing importance of Big Data and Artificial Intelligence (AI), this workshop served as a vital platform for participants to explore the application of machine learning in official statistics and monitoring the Sustainable Development Goals (SDGs). Experts from BPS-Statistics Indonesia, UNSD, and SIAP led training sessions covering core topics such as:

• Fundamentals of data science and machine learning
• Supervised and unsupervised classification methods
• Advanced regression techniques with hands-on practice using R
• Tree-based models, including random forests and decision trees
• Ethical considerations and identifying bias in machine learning models

Active Participation and Field Visit

The workshop was attended by 23 participants from National Statistics Offices (NSOs) across Asia-Pacific, including Indonesia, Malaysia, Kazakhstan, Timor-Leste, Thailand, Türkiye, and Vietnam.

On the first day, participants visited the BPS headquarters in Pasar Baru, Jakarta, where they were warmly welcomed by the Acting Deputy for Statistical Methodology and Information, Puji Ismartini, and her team. The visit began with a tour of the Integrated Statistics Center (PST), followed by a session at the BPS Command Center. After lunch, the group returned to Politeknik Statistika STIS to begin the training sessions.

Interactive Learning with Experts

The training was delivered by experienced mentors, including Christophe Bontemps from UNSIAP and Sean Lovell from UNSD. Participants learned foundational concepts and definitions in data science and machine learning, classification methods, and regression techniques. In addition to theoretical materials, mentors provided interactive demonstrations and hands-on exercises to deepen participants’ understanding.

Experts from BPS also contributed to the training. Lya Hulliyyatus Suadaa presented a case study on business classification at BPS, while Arie Wahyu Wijayanto discussed the application of the Random Forest algorithm in data analysis.

Strengthening Regional Collaboration

The event successfully reinforced collaboration among national statistics offices, with participants sharing experiences and best practices from their respective countries. Discussions also highlighted challenges in applying machine learning models, especially in terms of access to high-quality data and adapting models to varying institutional contexts.

Commitment to Future Innovation

As the workshop concluded, participants expressed strong commitment to applying the knowledge gained at their respective offices. Plans for further capacity-building initiatives and follow-up workshops were also discussed to ensure continued learning and innovation in leveraging machine learning for official statistics.

The success of this workshop marks an important step toward modernizing statistical systems through advanced data science and machine learning methodologies, paving the way for better data-driven decision-making across the Asia-Pacific region.