Abstract:
This project presents a smart crop suggestion and risk prediction system that leverages IoT and machine learning to assist farmers in making informed agricultural decisions. Using Arduino and ESP8266, real-time data is collected from soil moisture, pH, and DHT11 sensors to monitor key soil and environmental parameters. The data is stored in a MySQL database via PHP, while a weather API provides real-time climatic updates. A JavaScript-based Random Forest machine learning model analyzes one month of sensor data to predict the most suitable crops for the given conditions, assess the risks involved in growing specific crops, and suggest better alternatives if needed. Additionally, the system sends email alerts in the local language to farmers when adverse weather conditions are detected, enhancing proactive farming practices and reducing crop failure risks.
Introduction:
In modern agriculture, selecting the right crop based on soil and climatic conditions is essential for maximizing yield and ensuring sustainability. Farmers often rely on traditional knowledge and experience, which may not always align with changing weather patterns and soil conditions caused by environmental and human factors. This project aims to bridge that gap using technology by developing a smart system that can analyze real-time soil and climate data to provide accurate crop suggestions and risk assessments.
The core of this system is built on the Internet of Things (IoT), using sensors such as the DHT11 (for temperature and humidity), a soil moisture sensor, and a pH sensor. These sensors are interfaced with an Arduino and ESP8266 module, enabling wireless data transmission to a centralized database. This real-time data collection allows continuous monitoring of the soil environment, giving a more dynamic and responsive insight into farming conditions.
Alongside the sensor data, a weather API is integrated into the system to fetch live climate information, such as rainfall, temperature variations, and humidity levels. This allows the system to consider external environmental changes when making crop predictions. If any adverse weather conditions are detected, the system immediately sends alerts to the farmer via email, translated into the local language for easy understanding and timely response.
The software component is developed using PHP and MySQL for data storage and backend operations. For crop prediction and risk analysis, a machine learning model based on the Random Forest algorithm is implemented using JavaScript. This model is trained on one month of sensor data to predict the most suitable crops for cultivation under current conditions. It also assesses the level of risk involved in growing a particular crop and suggests more viable alternatives if the risk is high.
Overall, this project combines IoT, cloud computing, weather forecasting, and machine learning to create an intelligent, farmer-friendly system. By automating the crop selection process and providing timely alerts, it empowers farmers to make better agricultural decisions, reduce crop failures, and increase productivity, thereby contributing to smarter and more sustainable farming practices.

Objectives
• To monitor real-time soil parameters like moisture, temperature, humidity, and pH using sensors.
• To collect and store sensor data using Arduino, ESP8266, PHP, and MySQL.
• To integrate weather API for tracking live climate conditions.
• To send alerts to farmers via email in the local language during adverse weather.
• To use a JavaScript-based Random Forest model for crop prediction.
• To identify the most suitable crop based on environmental and soil data.
• To assess the risks involved in growing certain crops under current conditions.
• To suggest alternative crops if high risk is predicted.
• To support data-driven decision-making in agriculture.