Learning analytics refers to the use of data analysis techniques to uncover patterns in educational data for improving student outcomes and identifying areas where interventions are needed. A comparative case study approach was employed, involving data from administrative records and student surveys. Dropout rates were analysed using logistic regressions with robust standard errors to account for uncertainty in the estimates. Dropout rates showed a significant difference between schools (p < 0.05), indicating that learning analytics can be an effective tool for early intervention in dropout prevention, particularly when tailored to specific student needs and educational contexts. The study supports the use of learning analytics as a predictive tool in identifying students at risk of dropping out, which could inform targeted interventions by educators and policymakers. Schools should integrate learning analytics into their dropout prevention strategies, with ongoing monitoring to ensure continuous improvement based on data feedback. Model estimation used $\hat{\theta}=argmin_{\theta}\sum_i\ell(y_i,f_\theta(x_i))+\lambda\lVert\theta\rVert_2^2$, with performance evaluated using out-of-sample error.