Data Processing

Using NLP techniques (natural language processing), judict analyses vast amounts of legal documents so you don't have to. We thus simplify the complex legal research in EU Financial Law, by centralising data and by integrating the information into an easy-to-use database.

Through automated processes, we can guarantee daily updates of all content. Our last update was completed on 15/07/2024.

How we collect the data and stay up-to-date

Our legislative texts are retrieved directly from the website of the EU Publications Office (EUR-Lex) and synchronised with the EUR-Lex-Website every day. Through this, we can make sure to always provide you with the most up-to-date consolidated version of each legislative text. On top, we have implemented an algorithm-based system that detects all legislative amendments on a daily basis. With the help of this system, we are displaying a notice box at every article that is affected by an amending legal act published in the Official Journal but where the amendment has not yet been incorporated in the current EUR-Lex consolidation of the legislative text. This way we are safeguarding against the risk of you unintentionally working with an outdated consolidated version.

The related documents are sourced from the official websites of the European and national authorities through automated processes. On a daily basis, judict screens all relevant database entries, web pages and documents they provide. Whenever we detect new or updated documents, we add them to our document library. With our daily updating routine, we make sure that the library data is complete at all times.

For our document library, we screen the websites of the EBA, ESMA, EIOPA, ECB, BaFin and EUR-Lex. So far, we have analysed:

In total, this sums up to more than 40,000 analysed documents and web pages.

How we process the data

Once we have collected all web pages and documents, our algorithms decide on their relevance for the judict database. All web pages and documents that are classified as relevant are added. This is done through automated processes based on NLP techniques (natural language processing). Subsequently, the documents are mapped to the legal acts and the specific articles they relate to. We do this with algorithm-based pre-sorting followed by manual quality control. With this two-step process, we ensure the high quality of document relations in the judict database.

As regards the database, our algorithms detect:

Mapped to the specific articles of the respective legal acts, we provide a comprehensive overview of all the above mentioned resources.