Soil Moisture Monitoring Using 7 in 1 Sensor

## Dataset OverviewThis dataset contains a high-resolution, multi-parametric agricultural time-series capturing localized subsurface soil profiles over a continuous monitoring cycle between December 2024 and January 2025. The data was collected in Bavdhan, Pune, Maharashtra, India (Latitude: 18.5204° N, Longitude: 73.7744° E) at an elevation of 560 meters above sea level. The primary objective of this data collection is to provide a comprehensive, real-world empirical matrix to train machine learning models (such as Random Forest, Gradient Boosting, and SVM architectures) for precision crop recommendations, optimize smart irrigation protocols, and benchmark dynamic cyber-physical network routing and metaheuristic optimization algorithms . ## Data Collection and Hardware ConfigurationThe subsurface data collection architecture consists of:- **Sensor Node:** An industrial-grade, vacuum-encapsulated, corro hsion-resistant 7-in-1 Soil Multi-Parameter Probe. - **Microprocessing Unit:** An Arduino Uno edge microcontroller node.- **Protocol & Interface:** Data transmission was carried out via aalf-duplex RS485 physical interface executing the Modbus RTU serial communication protocol, stabilized using an automated software-serial arrangement to minimize line contention. ## Preprocessing & Data Cleaning PipelineDuring long-term field deployment under natural open-air conditions, environmental telemetry is inherently prone to high-frequency communication noise, packet drops, and transmission line timeouts. In this dataset's raw stream, these communication faults manifested as standard register-bound telemetry flags (e.g., `255` for 8-bit integer overflows, `25.5` for floating-point bounds, or `0.0` during initialization line-contention drops). ## Dataset Schema & Column DescriptionsThe processed dataset consists of a continuous tabular matrix with the following attributes: | Column Name | SI Unit | Data Type | Description || :--- | :---: | :--- | :--- || **timestamp** | DD-MM-YYYY HH:MM | DateTime String | Sequential temporal identifier mapping logging intervals. || **Moisture (%)** | % | Float | Volumetric Water Content (VWC) identifying soil moisture percentage. || **Temperature (°C)** | °C | Float | Subsurface profile thermal metrics measuring ambient soil energy. || **EC (us/cm)** | µS/cm | Integer | Apparent soil electrical conductivity tracking baseline salinity curves. || **pH** | - | Float | Logarithmic scale monitoring localized soil acidity-alkalinity profiles. || **Nitrogen (mg/kg)** | mg/kg | Integer | Available elemental Nitrogen (N) index mapping soil matrix fertility. || **Phosphorous (mg/kg)** | mg/kg | Integer | Available elemental Phosphorus (P) values indicating vital resource blocks. || **Potassium (mg/kg)** | mg/kg | Integer | Available elemental Potassium (K) metrics reflecting structural soil content. | ## Potential Applications- **Predictive Agronomy:** Building time-series deep learning architectures (LSTMs, GRUs) to forecast dynamic soil changes.- **Crop Suitability Analysis:** Evaluating machine learning classifiers to predict optimal seasonal crop cycles based on continuous N-P-K-pH distribution trends.- **Algorithm Benchmarking:** Utilizing multi-parametric real-world constraints to test local-optima avoidance and convergence profiles in evolutionary and swarm intelligence metaheuristics.