KU-MWQ: A Dataset for Monitoring Water Quality Using Digital Sensors

Overview

The KU-MWQ Dataset contains real-time water quality measurements collected from a fish pond at Khulna University, Bangladesh, using Arduino-based IoT sensors.

Data Collection

• Two sets of sensors deployed at 30 cm and 60 cm underwater depths for parallel data collection.
• Continuous 24-hour monitoring from January 15 to January 22, 2020 (7 days), with approximately 1 measurement per minute.
• Total of 9,623 timestamped records per depth, capturing day/night cycles and varying weather conditions (dry and rainy days).

Variables

• Timestamp: Date and time in YYYY-MM-DD [hh]:[mm]:[ss] format (24-hour).
• Temperature: Water temperature in °C (recorded at both depths).
• pH: pH value (recorded at 30 cm depth only).
• Turbidity: Water turbidity in NTU units (recorded at both depths).

Data Segments

• Day vs. Night: Daytime data (6:00 AM–6:00 PM) and nighttime data (6:00 PM–6:00 AM) are separately identified by row ranges.
• Weather conditions: Dry day data (rows 2–6148, 7462–9624) and rainy day data (rows 6149–7461).

Use Cases

• Determining environmental suitability for aquaculture and fish farming operations.
• Anomaly detection in aquatic environments using machine learning on time-series sensor data.
• Forecasting near-future water quality conditions for proactive pond management.