Official statistics
Methods and software for survey statistics, quality measurement, disclosure control, imputation, and statistical data workflows.
Statistics Austria
survey statistics
Statistics Austria - methods, software, and official statistics
Alexander Kowarik
Deputy Head of the Center for Methodology, Head of Statistical Methods, and Head of the Future Lab at Statistics Austria.
Statistician and open-source contributor focused on official statistics, reproducible methods, statistical computing, and practical tools for data quality.
Focus
Statistical methods that make it into production.
Alexander's work brings together official statistics, missing-data methodology, synthetic populations, disclosure control, survey estimation, and reliable open-source workflows for statistical production.
Computational statistics
Applied research on imputation, robust methods, calibration, simulation, and machine-learning-assisted statistical methods.
imputation
simulation
Open-source software
R packages and public repositories that make statistical methods reusable, transparent, and easier to apply in practice.
R
open methods
Current work
Methods for the next generation of official statistics.
Recent work connects classical statistical quality requirements with AI/ML, synthetic data, new data sources, and production-ready R workflows.
AI and machine learning for official statistics
Contributing to ESS work on responsible AI/ML methods, standards, training material, and use cases that can support statistical production.
AIML4OS
official statistics
Machine-learning-based imputation in VIM
Extending missing-data workflows with XGBoost and transformer-based methods while keeping the diagnostics and practical usability of VIM.
imputation
VIM
Synthetic data generation
Developing methods and practical guidance for synthetic data in official statistics, including links to disclosure control and reusable microdata workflows.
synthetic data
disclosure control
Web and mobile-network data
Working on quality, methodology, architecture, and survey-supported approaches for integrating new data sources into official statistics.
web data
MNO data
Software
Open packages and statistical infrastructure.
Collaborative contributions to open-source R packages that turn methodological work into reusable infrastructure for statistical production.
Missing data
VIM supports visual diagnostics and practical imputation workflows, from classical donor-based methods to newer machine-learning-based approaches.
missing data
R package
Disclosure control and synthetic data
Tools for protecting confidential statistical data, preparing public-use files, and simulating realistic synthetic populations.
disclosure control
synthetic microdata
Survey estimation
surveysd provides standard error estimation for complex household surveys, including bootstrap replicates, rotating panels, and flexible calibration.
bootstrap
panel surveys
Seasonal adjustment
x12 brings X12-ARIMA and X13-ARIMA-SEATS workflows into R, while persephone connects to the JDemetra+ ecosystem used in the European Statistical System.
seasonal adjustment
time series
Statistical infrastructure
Public R infrastructure around Statistics Austria methods work, including open-data access, production-oriented workflows, and maintained package ecosystems.
open data
R ecosystem
Research
Selected publications.
A selection of publications on imputation, disclosure control, synthetic data, and statistical software.
2025
Performance evaluation of machine learning-based imputation for missing data analysis in the R package VIM
Nina Niederhametner, Johannes Gussenbauer, and Alexander Kowarik, Statistical Journal of the IAOS.
2024
Simulation of Calibrated Complex Synthetic Population Data with XGBoost
Johannes Gussenbauer, Matthias Templ, Siro Fritzmann, and Alexander Kowarik, Algorithms.
2017
Simulation of synthetic complex data: The R package simPop
Matthias Templ, Bernhard Meindl, Alexander Kowarik, and Olivier Dupriez, Journal of Statistical Software.
2016
Imputation with the R Package VIM
Alexander Kowarik and Matthias Templ, Journal of Statistical Software.
2015
Statistical disclosure control for micro-data using the R package sdcMicro
Matthias Templ, Alexander Kowarik, and Bernhard Meindl, Journal of Statistical Software.
2014
Seasonal Adjustment with the R Packages x12 and x12GUI
Alexander Kowarik, Angelika Meraner, Matthias Templ, and Daniel Schopfhauser, Journal of Statistical Software.
2011
Iterative stepwise regression imputation using standard and robust methods
Matthias Templ, Alexander Kowarik, and Peter Filzmoser, Computational Statistics & Data Analysis.
Conferences
Conference contributions and presentations.
A selection of conference materials, with direct links to the public alexkowa/presentations repository.
Flexible machine-learning-based imputation with options for sequential imputation and predictive mean matching
Machine-learning-based imputation methods for the VIM ecosystem.
Open-source software for statistical methods in official statistics
Fourth Workshop on Methodologies for Official Statistics.
Building a statistics bot in almost no time
A practical workflow for building a statistics-focused assistant.
vimpute(): extending imputation methods in the VIM package
Extensions to the imputation methods available in the VIM package.
Can a traditional survey reduce the bias of mobile phone data based statistics?
Combining survey data with mobile-phone-based statistics to address bias.
Web Data for Official Statistics - Methodology, Quality, Production and Community
Session introduction and discussion on methodology, quality, production, and community.
DevOps for the statistical production process?
Modern software delivery practices for statistical production processes.
Teaching and community
Sharing methods through teaching, conferences, and open source.
Alongside methodological work at Statistics Austria, Alexander contributes to applied-statistics teaching and the international official-statistics software community.
Gerhart-Bruckmann-Preis 2024
Recipient of the Austrian Statistical Society's award for work that strengthens the public role and visibility of statistics.
recognition
statistics in public life
Teaching at JKU
Lecturer at the Institute of Applied Statistics for "Official Statistics" as part of the European Master in Official Statistics (EMOS), connecting practical statistical production experience with applied methodology and data quality.
EMOS
official statistics
Conference community
Regular contributions to uRos, ISI World Statistics Congress, WIN, and methodology workshops on R, production workflows, web data, imputation, and survey estimation. Co-chair of the organising committee for the UNECE Statistical Data Editing meeting series and member of the organising committee of the Use of R in Official Statistics. Member of the scientific committee of NTTS and co-organizer of the Statistical Scraping Interest Group.
uRos
SDE
NTTS
SSIG
Open-source methods
Contributing to public R packages used by national statistical institutes, with reproducible methods that can be inspected, reused, and improved by the community. Co-maintainer of the CRAN Task View for Official Statistics and Survey Statistics, and contributor to the awesome official statistics software collection.
R
open source
CRAN Task View
European projects
Collaboration across the European Statistical System.
Selected European projects where Alexander contributed to shared methods, quality frameworks, and reusable tools for modern official statistics.
Web Intelligence Network
ESS collaboration on web data for official statistics, including the Web Intelligence Hub, web-data quality, online job advertisements, and online enterprise characteristics.
Role: work package lead for quality, methodology, and architecture.
web data
ESS collaboration
ESSnet Big Data I and II
European Statistical System projects exploring how big data sources such as web data, mobile network data, and other new sources can move toward regular statistical production.
Role: led WPK on quality and methodology in ESSnet Big Data II.
big data
quality methods
ESSnet AIML4OS
One-stop-shop for artificial intelligence and machine learning in official statistics, developing shared capabilities, guidance, and use cases across the ESS.
Role: co-work package lead for synthetic data generation.
AI and ML
synthetic data
Centre of Excellence for Statistical Disclosure Control
European cooperation on statistical disclosure control, bringing national experts and Eurostat together to develop methods, guidance, and tools for protecting confidential statistical data.
Role: contribution to disclosure-control methods, guidance, and reusable tools.
confidentiality
SDC tools
MNO-MINDS
European project on methods for integrating mobile network operator data with other data sources for regular official statistics, including methodological guidance, training, and open-source tools.
Role: leading WP4 - proof-of-concept of an ad-hoc survey to improve MNO data.
MNO data
new data sources
Profiles and sources
Explore the live profiles.
Follow the profile, software, and presentation links for current publication details, source code, and conference materials.