This video visualizes the spatiotemporal evolution of predicted cockroach (Periplaneta americana) infestation levels across the urban sanitation network of Parque Paraíso Arenal, Córdoba, Spain, from April 7 to June 23, 2022. The heat maps are generated by a predictive machine learning model (primarily XGBoost) trained on real-time environmental data (CO₂, temperature, humidity) collected via an IoT sensor network deployed in eight strategically selected manholes.The visualization dynamically illustrates how risk zones shift and intensify over time, with warmer colors (red/orange) indicating areas with higher predicted probability of "High" infestation levels, based on elevated CO₂ concentrations (identified as the dominant bioindicator in the associated research). Cooler colors (green/blue) represent areas with low or no predicted activity ("None" or "Low").The dataset used to generate these predictions is available at https://doi.org/10.5281/zenodo.18022462.