Reliable access to analysis-ready Earth observation data and infrastructures for processing them has been a challenge with the increasing volumes and variety of data being generated daily through various Earth observation programmes. Recently, concepts centered around building cloud-native infrastructures that provide access to Earth observation data in efficient manners such as data cubes which facilitate rapid querying, filtering and retrieval have been garnering popularity. Moreover, efficient means of processing such vast volumes of data stored in data cubes through cloud computing frameworks such as Kubernetes are becoming more popular. This paper investigates the current state of the art techniques, methods and technologies used in cloud-native environments with a particular focus on the data cube initiative and "bring the user to the data" paradigm.

Monitoring air quality is critical for understanding pollution impacts on public health. The Copernicus Atmosphere Monitoring Service (CAMS) provides estimates of nitrogen oxides (NO_x), nitrogen dioxide (NO₂), fine particulate matter (PM_2.5) and inhalable coarse particles (PM₁₀), but its spatial resolution may limit accuracy at individual urban stations. This study introduces a graph centrality-based framework for station-level correction of CAMS European reanalysis data using in situ observations from ARPA Lombardia (2018 to 2021). Stations are modelled as nodes in a multiplex graph, with spatial influence quantified through six centrality measures including topology-based and context-aware variants such as the Adapted PageRank Algorithm (APA) and Eigendata. The centrality features are integrated with Geo-Temporally Weighted Regression (GTWR), Extreme Gradient Boosting (XGBoost) and Random Forest with Spatio-Temporal Kriging (RF-STK), evaluated under station, temporal and combined holdout strategies. Under station holdout, XGBoost achieves an $R^2$ of up to 0.66 for NO₂, while RF-STK can exceed this when residual spatial kriging is active. Adding OpenStreetMap (OSM)-context APA centrality yields pollutant-dependent changes: gains for NO_x and PM_2.5, mixed results for NO₂, and a temporal improvement for PM₁₀ with RF-STK. Results confirm that graph-theoretic structural context complements station-level CAMS correction but is not a general surrogate for pollutant dispersion.