Crop Recommendation System Leveraging Weather Forecast Data Using Random Forest

Agricultural decision-making has long depended on traditional practices and heuristics, which often fail to account for the variability in weather and soil conditions, leading to reduced yields and inefficient resource use. This study introduces a weather-aware Crop Recommendation System that employs the Random Forest machine learning algorithm to provide locationspecific crop suggestions. By integrating real-time meteorological data, including temperature, humidity, and rainfall, with soil nutrient profiles $(\mathrm{N}, \mathrm{P}, \mathrm{K})$, pH levels, and soil type, the system dynamically adapts to environmental changes and offers tailored recommendations. The model leverages both historical agricultural data and live weather feeds accessed via external APIs, allowing it to remain responsive to conditions such as unexpected droughts or rainfall. The proposed system is designed to identify the top five crops most suitable for the user's region and soil conditions. Extensive evaluation demonstrates a prediction accuracy of $\mathbf{9 9. 3 1 \%}$, confirming the system's robustness and effectiveness. This tool not only supports precision farming but also promotes sustainability by aiding farmers in optimizing input usage and maximizing productivity. A web-based interface makes the platform accessible to end users, particularly small-scale farmers seeking data-driven solutions.