Multivariate dataset from a low-cost IoT-instrumented recirculating aquaculture system (RAS) with synchronized operational logbook (Carassius auratus, 2025)

This dataset documents a low-cost, IoT-instrumented Recirculating Aquaculture System (RAS) deployed for monitoring an adult Carassius auratus specimen between 10 and 12 September 2025 in Colima, Mexico. The system, based on Raspberry Pi Pico W microcontrollers and a Firebase Realtime Database backend, captured 7,399 automatic records at 30-second intervals across four physicochemical variables: water temperature, pH, electrical conductivity, and ambient humidity, with an effective coverage of 95.7 % of the nominal sampling rate. Dissolved oxygen was measured manually using a colorimetric kit and is recorded separately (n = 17 measurements). Eleven operational events (sensor calibrations, dispenser activations, scheduled firmware resets, and maintenance interventions) are documented in a synchronized logbook, enabling contextualized labeling of outliers in the time series. A formal data dictionary is provided. The package is intended as infrastructure for downstream work on hybrid control models (rule-based + machine learning) in small-scale aquaculture, and as a reproducible empirical reference for the accompanying manuscript (in preparation for the Revista Iberoamericana de Automática e Informática Industrial).