The project investigates how changes and disruptions in financial variables propagate through network connections within individual markets and among markets in the region, with a special focus on the Croatian capital market. Market structure, the potential for forecasting market trends, and portfolio optimization are examined through three complementary research avenues: (1) Network analysis of capital markets, aimed at understanding the connectedness between stocks and markets in the region; (2) Development of predictive models for market trends based on machine learning and the integration of network metrics, sentiment, and macro indicators; (3) Portfolio optimization utilizing information on market topology to construct more efficient investment strategies. The focus is on the Croatian and regional capital markets, particularly under conditions of uncertainty and shocks. Data on stock prices and trading volumes, market sentiment, macroeconomic indicators, and other related features will be utilized. The project will result in scientific papers, an open analytical tool, and summaries of research insights intended for professional and institutional audiences. Emphasis is placed on the applicability of results and the early detection of financial instability. Through an interdisciplinary approach (economics, finance, mathematics), the inclusion of young researchers, and the integration of results into teaching, the project strengthens the research capacities of the Faculty of Economics and Business in Zagreb. It fosters teaching innovations, contributes to international visibility, and creates a foundation for future competitive grant applications.