Author(s) :

Tharun Mukesh.

Volume/Issue :

Abstract :

As fears of worldwide water shortage rise, finding sustainable solutions for water usage in agricultural fields is of utmost importance. The agricultural sector is the cause of approximately 70% of freshwater use globally, which is why the adaptation of techniques such crop rotation would potentially be pivotal in ensuring sustainable utility of the water consumption by encouraging soil health, water conservation and enhancing long-term resilience of the farmland. However, deducing appropriate crop rotations and optimal planting times is a challenge for farmers because it depends on a myriad of factors such as soil composition, rainfall patterns, temperature fluctuations and the life cycles of pests. This study investigated how data-driven tools can be used to aid improved crop rotation planning through the use of a broad set of agricultural and environmental variables. Based on a dataset from the FAO and World Data Bank of more than 28,000 entries spanning multiple countries and years, this project examined how rainfall, temperature, pesticide application, and other variables affect crop yield. The approach involved data processing by removing redundant columns, one-hot encoding of the categorical variables, numerical feature scaling by standardization and missing value handling. This was followed by the development of predictive regression models aimed at estimating crop yield and suggesting suitable crops and planting schedules. Model performance was measured in terms of standard metrics including mean absolute error, mean squared error, and R2. The tuned K-Nearest Neighbours model achieved the best performance with R2 = 0.99, MAE = 3257 hg/ha, and MSE = 78.5 million, showcasing high accuracy when predicting crop yield. The aim of this study was to help farmers seamlessly integrate crop rotation into their practice through a tool that provides straightforward insights to aid in maintaining soil health, controlling pest cycles, and reducing water consumption.