Development history
The system originated in the CNB’s Statistics Department, where it is still operated and developed with care and affection. More precisely, it is a product developed in-house.
Even for those of us who were there at the beginning, the following look back is an interesting journey.
1996
The first internal version, built in MS Access, is launched. Its main aim is to give CNB users a central database of time series spanning various statistical domains, complete with navigation, basic metadata and methodological information.
Users quickly take to the application. Installations multiply, even though each installation involves the ritual of visiting the user in person.
The internet in the Czech Republic is still in its infancy.
1997–1999
The data moves from MS Access to Oracle. We learn to work with a large database engine and live under the watchful eyes of strict database administrators.
We no longer have to deal with locked data files and slow queries. Nor do we need to visit users: the application is distributed as an installation package.
2000–2001
After surviving Y2K unscathed, we realise that the internet is suddenly all around us. We explore whether we can create a web application for retrieving data from ARAD.
In 2001, we launch a web application for internal users on the CNB intranet. No more installations: a web browser is enough.
2002–2005
In 2002, a public web application is launched on the CNB website. The internet changes the accessibility of data and information.
We watch usage figures rise and begin staffing the service around major data releases. ARAD becomes the CNB’s main platform for publishing statistical data. We gradually move away from publishing dozens of MS Excel files on the CNB website. We also start supplying international financial authorities with data from ARAD, bringing publication and reporting processes together.
2006–2022
In December 2006, a new version launches on the CNB website, modernising the interface and adding data news to keep users informed. With some determination, the system can even be used on PDAs and early smartphones — the iPhone does not yet exist.
In the following years, we focus on processes that ensure timely and accurate data publication. Gradually, almost everything is versioned and controlled by timestamps. The system remains simple on the surface while becoming more complex behind the scenes.
The visible interface stays the same. Time passes. For 16 long years, ARAD’s appearance remains unchanged while its content steadily expands.
2023–2025
In January 2023, after 16 years, we launch a new version. We are full of anticipation but cautious. The original version stays in operation until summer 2023, giving long-standing users time to get used to the new system and reconnect their data processes.
Further updates gradually extend the API and improve individual features.
2026–present
After three years of operating the new generation of ARAD, we find that the work is still not finished. Users keep showing us what works, what could work better and where we may have made their work less easy than we originally thought.
We are therefore preparing another major update. We are changing how indicators are browsed and filtered, reworking database search and reorganising sets. We are also adding API Builder, which lets users construct API calls without first studying every parameter.
Many other changes are less visible but all the more practical. The aim is to help users find their way around, locate the data they need faster and spend less time wondering where we have hidden a feature.
Even after thirty years, ARAD continues to evolve. Technology, ways of working with data and user expectations change. The basic goal remains the same: to make the CNB’s statistical data accessible, clear and reliable, all in one place.