Chronic Kidney Disease (CKD) poses a major public health burden in Jalisco, Mexico. However, data limitations often hinder comprehensive analysis. Leveraging diverse public data sources through multimodal approaches offers a promising path to deeper insights: in this study, we propose a multimodal machine learning workflow, integrating public mortality and geospatial data to identify demographic and geographic factors associated with CKD, using Explainable Artificial Intelligence for interpretation. We combine 103,142 tabular mortality records and two geopackages of spatial data. After extensive preprocessing and feature engineering, we train several classifiers using hyperparameter optimization. We evaluate the performance of the models using the Area Under the Precision-Recall Curve with stratified 5-fold cross-validation. We then employ Explainability methods to interpret the top-performing models. Tree-based models, particularly XGBoost, showed the best performance, albeit with modest predictive capability (AUPRC=0.255) due to data constraints. Explainable Artificial Intelligence (XAI) consistently highlighted age, health affiliation, employment status, and postal code-derived regions as key predictive features. While predictive power is still limited, integrating public mortality and geospatial data using multimodal machine learning proved feasible. Our approach combining multimodal machine learning and explainability effectively identified potential risk factors, underscoring the need for richer clinical and environmental data for improved Chronic Kidney Disease prediction in Jalisco.