Eurostat Universal Explorer: An Interactive Web Application for Statistical Data Visualization with AI-Assisted Analysis

This application was developed as part of a dissertation project, motivated by direct professional experience in the field of statistics and by the growing need to make statistical data more accessible in today’s data-driven context. Despite the wealth of information published by Eurostat-collected from all European Union member states and freely available through a public API-retrieving and interpreting this data in practice remains difficult for anyone without a technical background. Eurostat Universal Explorer was built to close this gap. The application fetches live data directly from the Eurostat REST API and presents it through interactive charts and maps, without requiring the user to understand the underlying data format. An AI module powered by GPT-4o-mini translates the raw statistics into plain-language explanations accessible to any user, regardless of their expertise. One challenge encountered during development was that the AI would occasionally misstate the direction of a trend-describing a decline as an increase. This was resolved through iterative prompt engineering, with four explicit numerical verification rules added to the system prompt. The application was validated on six real datasets covering unemployment, GDP per capita, CO2 emissions, life expectancy, early school leaving, and inflation across EU member states. The application is publicly deployed on Render and Netlify.