This thesis explores Big Data-based approaches for understanding and preparing to forecast PM10 air pollution levels. The study focuses on data collection and preprocessing, followed by correlation analysis and heatmap visualization to reveal key relationships. A specific emphasis is placed on the interaction between wind speed and PM10 levels on the previous day. Principal Component Analysis (PCA) is applied to reduce dimensionality, and a preliminary neural network analysis is conducted using winter hourly data. While predictive modeling is left as future work, this research lays the groundwork for more advanced forecasting solutions. The findings support the potential of data-driven methods in environmental air quality analysis.