Data relevant to consumer policy is no longer generated solely in controlled survey settings – it is also available in large quantities in digital spaces, such as social media, news portals, forums, product reviews and digital communication platforms, although access may vary depending on the platform and data source. Computer- and AI-assisted analytical methods make it possible to systematically access, structure and evaluate these digital data sources.

For our clients, we use computer- and AI-assisted methods to download (web scraping) and efficiently analyse large volumes of unstructured text, image or interaction data. The analytical methods include natural language processing (NLP) techniques for automated classification, topic modelling and sentiment analysis, as well as machine learning techniques. These methods can be used both to evaluate our own survey data – such as open-ended responses from questionnaires or interview transcripts – and to analyse publicly available digital sources. Human, scientific interpretation and quality assurance by our team remain central to this process at all times.