How BERD@NFDI makes research data more accessible
From AI training to knowledge graphs: new services and infrastructure are helping researchers to analyse and link complex data more efficiently and make it usable in the long term.
What this article is about:
- how BERD@NFDI supports researchers in the economic and social sciences,
- what new training opportunities are emerging in the fields of AI and research data, and
- how new data infrastructures are helping to make complex research data more accessible.
Research data in the economic and social sciences is becoming increasingly voluminous and complex. Studies, company data, texts, surveys and digital platform data must be stored, documented, analysed and made available for long-term use.
This is precisely where the NFDI consortium BERD@NFDI comes in. Its aim is to support researchers with suitable tools, training programmes and technical infrastructure – so that research data can be used securely, transparently and efficiently.
To this end, BERD@NFDI will further expand its community and training programmes in 2026 – both in terms of technology and content.
New training programmes on AI and data skills
With the expanded BERD Academy and additional training modules such as ASSURED, new opportunities for further training are emerging in areas such as:
- Data Science,
- data ethics,
- legal issues relating to data handling, and
- methods of artificial intelligence.
This also includes what is known as reinforcement learning – a field of AI in which systems learn to make better decisions through experience.
Reinforcement Learning Explained in Simple Terms
How do machines actually learn to make decisions? That is precisely what reinforcement learning is all about.
In this process, a system ‘learns’ step by step through feedback from its environment: good decisions are rewarded, whilst poor decisions are discarded. In this way, its behaviour improves autonomously over time.
Such methods are also increasingly being used in economics and the social sciences – for example, for forecasting, adaptive experiments or automated decision-making processes.
New infrastructure for complex research data
As well as offering training programmes, BERD@NFDI is also working on technical solutions for research data.
This is because much of the data is not neatly structured, but is scattered across documents, spreadsheets, texts or various databases. As a result, research often becomes time-consuming and confusing.
New data infrastructures are intended to help link such information more effectively and make it usable.
Knowledge graphs: connections rather than isolated data
So-called knowledge graphs play an important role in this.
They function like intelligent knowledge networks: information is not merely stored, but linked together. This makes it possible to identify connections more quickly and analyse data from different sources collectively.
This makes it easier, for example, to:
- searching for relevant information,
- linking different data sets, and
- automated analysis of large volumes of data.
In this way, individual data points are transformed into interconnected and machine-readable knowledge structures.
About BERD@NFDI’s knowledge graph infrastructure
Example: The AKF (Aktienführer) Knowledge Graph is freely accessible and open for use. It contains structured (meta)data on German listed companies from the Hoppenstedt Share Directory for the period from 1956 to 2018. The metadata includes official company names with their respective periods of validity, company identifiers such as ISIN and WKN, and the corresponding Aktienführer IDs.
Contact the Research Data Management team
Annette Strauch-Davey , M. A.
Scientific representatives with teaching duties
Faculty of Health | Research data management
Alfred-Herrhausen-Straße 50
58455 Witten