Smart Farming IoT Sensor Dataset for Crop Yield Prediction

This dataset contains structured smart farming records including soil moisture, temperature, humidity, rainfall, sunlight exposure, fertilizer usage, crop type, and yield outcomes. The dataset has been processed through a structured exploratory data analysis workflow including duplicate removal, missing value treatment, sensor value range validation, timestamp normalization, and categorical standardization. Statistical anomaly detection techniques were applied to identify unrealistic sensor spikes and inconsistent crop entries. The final dataset is exported in a clean CSV format optimized for regression modeling, classification, and precision agriculture research.

Data Quality

The dataset is organized in a normalized tabular format with consistent column naming and validated numeric data types. Duplicate sensor entries have been removed to preserve integrity. Environmental variables such as soil moisture, temperature, and humidity have undergone range validation to eliminate physically unrealistic values. Categorical attributes including crop type and fertilizer category have been standardized. Timestamps were normalized to maintain chronological continuity. The dataset is analytics-ready and supports machine learning workflows for yield forecasting.

Use Cases

This dataset can be used for crop yield prediction modeling, precision irrigation planning, and smart farming analytics. Data scientists can build regression and machine learning models to forecast yield based on environmental sensor data. Agri-tech researchers can analyze the relationship between soil conditions, weather patterns, and crop productivity. Policymakers and precision agriculture innovators can use it for sustainable farming research and resource optimization. The dataset also supports academic research in IoT-driven agriculture and climate-smart farming.

Limitations

The dataset is community-sourced and may reflect simulated or field-specific IoT sensor readings rather than universal agricultural conditions. Sensor accuracy depends on the original data collection setup. It may not fully represent all agro-climatic zones. The dataset is intended for research and analytical purposes and should not replace official agronomic recommendations.

Methodology

The dataset was prepared using a structured validation framework including schema standardization, duplicate removal, sensor value range validation, missing value treatment, and timestamp normalization. Statistical anomaly detection was applied to identify unrealistic sensor readings and inconsistent categorical entries. Logical consistency checks were performed across environmental and crop-related variables. The final dataset was exported in UTF-8 encoded CSV format to ensure compatibility with IoT analytics tools, machine learning libraries, and business intelligence platforms.