🌾 Crop Recommendation with Machine Learning

This mini-project demonstrates how Machine Learning can assist farmers in choosing the most suitable crop for their field based on soil metrics. The goal is to predict the best crop using a simple dataset containing soil nutrient levels and pH values.

---

🧠 Project Objective

To build a multi-class classification model that can predict the optimal crop for a field using the following features:

• N: Nitrogen content in the soil
• P: Phosphorous content in the soil
• K: Potassium content in the soil
• pH: Acidity level of the soil

We also identify the single most important feature contributing to the prediction using individual logistic regression models.

---

📁 Dataset

Filename: soil_measures.csv
Each row in the dataset represents soil conditions of a field and the optimal crop to grow.

Columns:

• N: Nitrogen ratio
• P: Phosphorous ratio
• K: Potassium ratio
• ph: pH value of the soil
• crop: Target variable (crop name)

---

⚙️ Model & Approach

• Algorithm Used: Logistic Regression (Multinomial)
• Task: Multi-class Classification
• Evaluation Metric: Weighted F1 Score

We trained separate models for each feature to evaluate their individual contribution to prediction accuracy.

---

📊 Results

Feature	F1 Score
N	0.091
P	0.148
K	0.239 ✅
pH	0.045

👉 Potassium (K) turned out to be the best single feature for predicting crop type.

---

📝 Key Learnings

• Even with minimal features, machine learning models can provide useful insights for agriculture.
• Among the soil metrics, Potassium (K) had the highest predictive power in isolation.
• Further improvement can be done using all features together, hyperparameter tuning, and advanced classifiers (e.g., Random Forest, XGBoost).

---

💡 Future Work

• Build a full pipeline using all features
• Apply advanced models like Decision Trees or Random Forests
• Add feature importance analysis using SHAP or permutation importance
• Integrate real-time soil sensor data (IoT)

---

📌 Requirements

• Python 3.x
• pandas
• scikit-learn

Install required packages:

pip install pandas scikit-learn

📂 Usage

import pandas as pd
from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn import metrics

Make sure soil_measures.csv is in the same directory as your script or Jupyter notebook.