IST (Integrated System for data collection and processing) is a comprehensive platform fully developed by the Statistical Office of the Republic of Serbia (SORS)
to modernize and simplify statistical production. At its core, IST is a metadata-driven environment
designed to unify and automate the workflows of collection, processing,
output tables generation, and reporting (on survey progress and key Quality Indicators)
in alignment with the GSBPM model.
The initial concept was introduced in 2006.
The platform was built around a simple yet powerful idea:
If data is stored in databases, why not store the application logic in a database as well?
The IST platform is built on two core components:
- Metadata repository — the IST Metadata Database, which defines every aspect of a statistical application, including variables, forms, logic, validations, and more.
- Metadata interpreter — the IST Interpreter, a dynamic engine that reads these metadata definitions and, in real time, generates and executes:
- Data entry interfaces
- Validation and control logic
- Data editing environments
- Advanced search functions (including cross-survey linking)
- Reporting and processing procedures
To ensure cross-platform compatibility, IST provides three separate interpreters, all of which operate on the same central metadata database:
- An interpreter for Windows-based environments (desktop, Windows laptop CAPI, CATI),
- A web interpreter for browser-based (CAWI) applications, and
- An Android interpreter for mobile (CAPI) data collection.
This design guarantees consistency across platforms while allowing applications to be executed in diverse operating environments.
In the following sections, we will refer to all three interpreters collectively as the IST Interpreter.
It is important to note that IST itself does not store or retain statistical data.
All individual and aggregated data related to surveys are stored in separate relational databases,
which may reside on one or multiple servers within the NSO infrastructure.
By adopting this model, IST achieves true data and application integration. All information remains in one place, consolidated within a single,
user-friendly platform — empowering statistical offices to operate more efficiently, adapt more rapidly,
and maintain full control over their data lifecycle.
During the time, IST has evolved through continuous enhancement and international collaboration.
It was created to address the growing complexity and fragmentation in statistical systems, particularly in environments using siloed, survey-specific tools
that lacked standardization. By centralizing metadata and automating application generation,
IST reduces dependence on repetitive programming and improves consistency across statistical domains.
What sets IST apart is its philosophy as a free metadata driven platform made available to other national statistical offices.
Rather than a commercial product, IST is a government-developed digital public good, shared with the goal of enabling peer institutions
to accelerate digital transformation without incurring high licensing or vendor-lock-in costs.
The system's flexibility is demonstrated by its successful deployment across various countries and contexts.
Today, IST is actively used in national statistical offices such as those of
Albania, Bosnia and Herzegovina, Kyrgyzstan, Montenegro, Republika Srpska, and Serbia.
These institutions use IST to manage a broad range of statistical operations,
including Population and Agriculture Censuses, LFS, HBS, CPI, SILC, and many others,
often customizing their implementations through the system’s powerful developer tools and modular architecture.
Importantly, this collaborative approach brings mutual benefits. While other NSOs benefit from a proven, standards-compliant platform,
SORS gains from extending the IST development community and continuously learning from partners’ experiences, requirements, and innovations.
This exchange of knowledge strengthens the quality, relevance, and scalability of IST for all users.
IST’s interpreter architecture consists of key modules such as: