Abstract
This study presents the development and deployment of a low-cost, multi-sensor system, integrated with IoT technology, designed for water monitoring and forecasting in the Cabuyao City area, a tributary to Laguna de Bay. The system tracks six key physicochemical parameters: temperature, dissolved oxygen, pH, oxidation-reduction potential, turbidity, and total dissolved solids using low-cost, field-validated sensors connected to ESP32 microcontrollers and LoRa modules. These parameters are sent to a Django-PostgreSQL backend and visualized through a ReactJS dashboard, which displays microbiological and chemical data from third-party validation. We deployed the system at two monitoring stations and collected over 1,000 records. For forecasting, the researcher trained CNN-BiLSTM models on both historical and real-time data, achieving a high R² of 0.91. This deep learning framework outperformed ARIMA, yielding better results in MAE and RMSE metrics. Additionally, the framework employed the water quality index to provide monthly assessments of the lake's environmental status. The framework offers a scalable and energy-efficient solution that facilitates timely water governance. It supports Sustainable Development Goals (SDGs) 6 and 13 by enhancing community engagement and driving decision-making through data and innovative low-carbon monitoring.
