ISTARI delivered a real-time time-series analysis solution for the German Federal Ministry of Economic Affairs using AI-powered web monitoring, NLP, and survey integration. This framework enabled early detection of sector-specific COVID-19 impacts to support targeted and efficient policy responses.
Delayed data and slow reporting in the face of a fast-moving crisis
When COVID-19 hit, the German Federal Ministry for Economic Affairs and Energy faced a serious information gap. Conventional data sources — such as economic indicators, trade statistics, and industry reports — were too slow and too aggregated to reflect the urgent and uneven impact of the crisis across sectors. To ensure targeted, effective economic intervention, policymakers required real-time, high-resolution insights about firms’ operational status, financial risk, and sector-specific vulnerabilities.
Real-time web surveillance with time-series analysis, NLP, and data fusion
ISTARI developed a novel time-series analysis framework that combined real-time web signals with firm-level survey data and credit ratings. In phase one, we analyzed over 1 million German company websites using advanced Natural Language Processing (NLP) to detect COVID-19 references. Our models automatically classified mentions across different semantic contexts — such as "operational disruptions," "adaptation measures," or "general communication." These signals offered early, sector-specific insights into how industries were affected.
In parallel, the Ministry conducted targeted business surveys to gather firm-level responses. We then integrated credit rating data as a third pillar to assess companies’ financial health over time. By linking early web indicators with future credit scores, we validated the framework's predictive capability — early digital signals anticipated changes in solvency as measured a year later.
To find out more, read the open-access study
Faster response times, better targeting, and minimized fiscal waste
With near real-time monitoring updated regularly, the Ministry gained crucial early insights into the sectors most affected by the pandemic. This data allowed for prioritization of support programs, mitigation policies, and emergency relief funding based on actual need rather than lagging indicators. The approach significantly improved both speed and precision in economic decision-making and minimized fiscal inefficiencies by avoiding untargeted support. ISTARI’s webAI-powered model not only informed immediate crisis decisions — it also established a replicable blueprint for future real-time economic monitoring.
