AAR MAHAVEER ENGINEERING COLLEGE
SmartFarm : Data-Driven Crop Recommendation System
UNDER THE GUIDANCE: PRESENTED BY:
MRS. JIVERIA JABEEN SADIA BEGUM (218P1A0528)
NAZIYA BEGUM(218P1A0522)
CH. PURNIMA (218P1A0507)
D. SHARDHA (218P1A0509)
TABLE OF CONTENTS
1.Abstract
2.Introduction
3.Problem statement
4. Objective
5.Literature Survey
Existing System
Proposed methods
6.Modules
7.Conclusion
ABSTRACT
Efficient crop selection based on soil health is crucial for sustainable
agricultural productivity. This project proposes a Crop
Recommendation System that utilizes essential soil nutrient values—
namely nitrogen (N), phosphorus (P), potassium (K), along with pH,
temperature, humidity, and rainfall—as input parameters to
recommend the most suitable crop for a given piece of land.
Leveraging machine learning algorithms such as Random Forest and
Decision Trees, the system is trained on a dataset of soil and crop
characteristics to generate accurate recommendations. By analyzing
the nutrient composition and environmental conditions, the model
helps farmers make informed decisions, optimize fertilizer usage,
and enhance crop yield. The system is designed to be user-friendly.
INTRODUCTION
• Agriculture is the backbone of many economies, and selecting the right crop
Based on soil and climate conditions is essential for maximizing productivity and
sustainability.
• Traditional crop selection often relies on experience or trial-and-error, which
can lead to Poor yield, soil degradation, and inefficient use of resources.
• A Crop Recommendation System uses data such as nitrogen (N),
phosphorus(P), potassium (K), pH, temperature, humidity, and rainfall to suggest
the most suitable crop for a specific region or plot.
• Machine learning algorithms analyze historical and real-time data to identify
patterns and recommend crops with the highest chances of success.
• This system empowers farmers with actionable insights, supports precision
farming, and contributes to food security by improving crop planning and
resource efficiency.
PROBLEM STATEMENT:
Agricultural productivity is significantly influenced by factors such as
soil quality, climatic conditions, and crop selection. However, farmers
often rely on traditional knowledge or generalized advice, which may
not align with modern farming needs. This can lead to:
Poor Crop Yields due to unsuitable crop choices.
Soil Degradation from repeated incorrect crop cycles.
Resource Misuse (excess water, fertilizers, and pesticides).
Financial Losses from low productivity and poor market demand.
There is a growing need for an intelligent, data-driven system that can
assist farmers in making informed decisions by recommending the
most suitable crops based on real-time environmental and market
data.
Objectives: Building a Smart Crop
Recommendation System
To develop a machine learning-based model that accurately recommends
suitable crops based on soil nutrients (N, P, K), pH, temperature, humidity,
and rainfall.
To assist farmers in making data-driven decisions that improve crop
yield and reduce the risk of crop failure.
To promote efficient use of fertilizers and resources by aligning crop
selection with soil health and environmental conditions
To support sustainable and precision agriculture by minimizing
environmental impact and maximizing productivity through informed
crop planning
Literature Survey: Existing Approaches and Challenges
Existing Systems ML Algorithms Research Gaps Precision Focus
Many crop recommendation Commonly used models: Need for better data quality Data-driven farming improving
tools exist with mixed success. Random Forest, SVM, neural and context-aware predictions. decision accuracy and
networks. efficiency.
EXISTING SYSTEM
1. Traditional Methods
•Farmers rely on personal experience, seasonal trends, or advice from local
experts to decide which crops to grow.
•These methods often ignore precise soil nutrient levels and climatic variations,
leading to suboptimal yield.
2. Government Agricultural Portals
•Platforms like India’s Krishi Vigyan Kendra (KVK) and Soil Health Card
Scheme provide soil reports and general crop advice.
•However, recommendations are often broad, not personalized, and lack
integration with AI or machine learning technologies.
3. Mobile Apps (e.g., Kisan Suvidha, RMLTraditional Methods
• AgTech)
•Offer basic crop advisory, weather forecasts, and market prices.
•While useful, they typically do not provide crop recommendations based on
specific soil nutrient values (N, P, K) and pH.
4. Academic/Prototype Projects
•Some research papers and university projects have implemented machine
learning-based crop recommendation systems.
•These are often limited to small datasets, not deployed at scale, and may lack
user-friendly interfaces.
PROPOSED SYSTEM
The system takes critical soil nutrients (N, P, K), pH level, and environmental
conditions (temperature, humidity, rainfall) as input to assess soil and climatic
suitability.
• Machine Learning Integration: It uses trained ML models (e.g., Random Forest,
Decision Tree) to analyze input data and predict the most suitable crops based on
historical patterns and optimal growing conditions.
• Data Preprocessing & Validation: Raw input values are cleaned, normalized, and
checked to ensure accuracy before being fed into the model for prediction.
• User Interface: Farmers or users can easily enter soil and weather data through a
simple interface (web or mobile app) and receive crop recommendations instantly.
• Output & Benefits: The system outputs the best crop(s) to cultivate under the given
conditions, helping improve yield, reduce input waste, and promote precision farming.
Machine Learning Model Selection
Algorithm Usage
Random Forest Strong accuracy, handles diverse
features
Support Vector Machine (SVM) Effective in classification with clear
margins
Neural Networks Model complex relationships, deep
learning
Models trained on historical data and validated with cross-validation metrics like
accuracy and F1-score.
Data Acquisition and Preprocessing
Data Sources Data Types Preprocessing
• Weather stations • Temperature, humidity, rainfall • Handling missing data
• Soil sensors • Soil pH, nutrient levels (NPK) • Outlier detection and removal
• Historical yield records • Normalization and feature engineering
Modules and Components
1. Data Collection Module
Gathers input data such as nitrogen (N), phosphorus (P), potassium (K), pH,
temperature, humidity, and rainfall from user input or sensors/datasets.
2. Data Preprocessing Module
Cleans, normalizes, and formats the raw data for effective use in the machine
learning model; handles missing or inconsistent values.
3. Machine Learning Model Module
Uses algorithms like Random Forest, Decision Tree, or SVM to train on historical
crop-soil data and predict the best-suited crop for the input conditions.
4. Crop Recommendation Engine
Generates crop suggestions based on the model’s output and displays the most
suitable crop(s) for the given inputs.
5. User Interface Module
Provides an intuitive web or mobile-based platform for users to input data and
view recommendations; may include features like multilingual support or voice
input.
Conclusion and Future Directions
Success Scalability Future Research Vision
Data-driven recommendations System adaptable for Include climate change models AI-driven innovations for
enhance farming outcomes. widespread agricultural use. and pest control integration. sustainable global food
security.
THANK YOU
