import gradio as gr
import pandas as pd
import numpy as np
from sklearn.ensemble import RandomForestClassifier
import matplotlib.pyplot as plt
import seaborn as sns
from datetime import datetime
# Synthetic Dataset Creation
def create_synthetic_dataset():
# Districts of Andhra Pradesh
districts = [
'Anantapur', 'Chittoor', 'East Godavari', 'Guntur', 'Krishna',
'Kurnool', 'Nellore', 'Prakasam', 'Srikakulam', 'Visakhapatnam',
'Vizianagaram', 'West Godavari', 'YSR Kadapa'
]
# Common crops in Andhra Pradesh (including new crops)
crops = [
'Rice', 'Maize', 'Corn', 'Cotton', 'Groundnut', 'Red Gram (Toor Dal)',
'Green Gram (Moong Dal)', 'Black Gram (Urad Dal)', 'Sugarcane',
'Chilli', 'Pepper', 'Turmeric', 'Tobacco', 'Sweet Potato', 'Mango',
'Banana', 'Coconut', 'Cashew', 'Soybean', 'Sunflower',
'Jowar (Sorghum)', 'Bajra (Pearl Millet)'
]
# Months
months = ['January', 'February', 'March', 'April', 'May', 'June',
'July', 'August', 'September', 'October', 'November', 'December']
# Create synthetic data
np.random.seed(42)
num_samples = 5000
data = {
'District': np.random.choice(districts, num_samples),
'Month': np.random.choice(months, num_samples),
'Temperature': np.random.uniform(20, 40, num_samples),
'Rainfall': np.random.uniform(0, 300, num_samples),
'Soil_Type': np.random.choice(['Black', 'Red', 'Alluvial', 'Laterite'], num_samples),
'Crop': np.random.choice(crops, num_samples),
'Suitability': np.random.choice([0, 1], num_samples, p=[0.3, 0.7])
}
# Add some logical patterns based on real-world knowledge
for i in range(num_samples):
district = data['District'][i]
month = data['Month'][i]
# Adjust temperature based on month
if month in ['December', 'January', 'February']:
data['Temperature'][i] = np.random.uniform(15, 28)
elif month in ['March', 'April', 'May']:
data['Temperature'][i] = np.random.uniform(28, 42)
else:
data['Temperature'][i] = np.random.uniform(25, 35)
# Adjust rainfall based on district and month
if district in ['Visakhapatnam', 'Srikakulam', 'Vizianagaram']:
if month in ['July', 'August', 'September']:
data['Rainfall'][i] = np.random.uniform(150, 300)
else:
data['Rainfall'][i] = np.random.uniform(50, 150)
elif district in ['Anantapur', 'Kurnool', 'YSR Kadapa']:
data['Rainfall'][i] = np.random.uniform(0, 100)
else:
if month in ['July', 'August', 'September']:
data['Rainfall'][i] = np.random.uniform(100, 250)
else:
data['Rainfall'][i] = np.random.uniform(20, 100)
# Adjust suitability based on some logical conditions
crop = data['Crop'][i]
# Rice needs more water
if crop == 'Rice' and data['Rainfall'][i] < 100:
data['Suitability'][i] = 0
# Corn needs moderate water and warm temperature
if crop == 'Corn' and (data['Rainfall'][i] < 50 or data['Temperature'][i] < 20):
data['Suitability'][i] = 0
# Pepper needs warm, humid conditions
if crop == 'Pepper' and (data['Temperature'][i] < 20 or data['Rainfall'][i] < 100):
data['Suitability'][i] = 0
# Sweet Potato grows well in warm conditions with moderate rainfall
if crop == 'Sweet Potato' and (data['Temperature'][i] < 20 or data['Rainfall'][i] > 250):
data['Suitability'][i] = 0
# Groundnut grows well in Anantapur
if crop == 'Groundnut' and district == 'Anantapur':
data['Suitability'][i] = 1
# Coconut grows well in coastal areas
if crop == 'Coconut' and district in ['East Godavari', 'West Godavari', 'Visakhapatnam']:
data['Suitability'][i] = 1
# Chilli grows well in Guntur
if crop == 'Chilli' and district == 'Guntur':
data['Suitability'][i] = 1
# Corn grows well in Krishna and Guntur
if crop == 'Corn' and district in ['Krishna', 'Guntur', 'West Godavari']:
data['Suitability'][i] = 1
# Pepper grows well in coastal and hilly areas
if crop == 'Pepper' and district in ['Visakhapatnam', 'Srikakulam', 'Vizianagaram']:
data['Suitability'][i] = 1
# Sweet Potato grows well in various districts
if crop == 'Sweet Potato' and district in ['East Godavari', 'West Godavari', 'Krishna']:
data['Suitability'][i] = 1
df = pd.DataFrame(data)
return df, crops, districts, months
# Create dataset
df, crops, districts, months = create_synthetic_dataset()
# Train machine learning model
def train_model(df):
# Convert categorical variables to numerical
df_encoded = pd.get_dummies(df, columns=['District', 'Month', 'Soil_Type', 'Crop'])
X = df_encoded.drop('Suitability', axis=1)
y = df_encoded['Suitability']
model = RandomForestClassifier(n_estimators=100, random_state=42)
model.fit(X, y)
return model
model = train_model(df)
# Crop information and precautions (updated with new crops)
crop_info = {
'Rice': {
'description': 'Staple food crop requiring abundant water',
'precautions': [
'Ensure proper water management (5-10 cm standing water)',
'Use certified seeds for better yield',
'Control weeds in early stages',
'Monitor for pests like stem borers and leaf folders'
]
},
'Maize': {
'description': 'Versatile cereal crop grown in diverse conditions',
'precautions': [
'Plant in well-drained soil',
'Maintain proper spacing (60x20 cm)',
'Apply nitrogen in split doses',
'Watch for fall armyworm infestation'
]
},
'Corn': {
'description': 'Sweet corn variety popular for direct consumption',
'precautions': [
'Plant in well-drained loamy soil',
'Maintain spacing of 60x25 cm',
'Harvest when silks turn brown and dry',
'Control corn earworm and aphids',
'Irrigate regularly during grain filling stage'
]
},
'Cotton': {
'description': 'Important cash crop known as "white gold"',
'precautions': [
'Use Bt cotton seeds for pest resistance',
'Monitor for pink bollworm',
'Practice crop rotation to prevent soil depletion',
'Avoid waterlogging in fields'
]
},
'Groundnut': {
'description': 'Oilseed crop important for protein and oil',
'precautions': [
'Plant in well-drained sandy loam soil',
'Apply gypsum at flowering stage',
'Harvest at proper maturity to avoid aflatoxin',
'Store in dry conditions'
]
},
'Red Gram (Toor Dal)': {
'description': 'Important pulse crop rich in protein',
'precautions': [
'Drought resistant but needs irrigation at flowering',
'Treat seeds with rhizobium culture',
'Control pod borer with recommended pesticides',
'Harvest when 80% pods are mature'
]
},
'Green Gram (Moong Dal)': {
'description': 'Short duration pulse crop',
'precautions': [
'Grows well in well-drained soils',
'Short duration (60-70 days)',
'Susceptible to yellow mosaic virus - use resistant varieties',
'Harvest when 80% pods are mature'
]
},
'Black Gram (Urad Dal)': {
'description': 'Important pulse crop for protein',
'precautions': [
'Grows well in black cotton soils',
'Treat seeds with rhizobium culture',
'Control leaf spot diseases',
'Harvest when pods turn black'
]
},
'Sugarcane': {
'description': 'Important cash crop for sugar production',
'precautions': [
'Requires heavy irrigation',
'Use disease-free setts for planting',
'Control early shoot borer',
'Harvest at proper maturity (10-12 months)'
]
},
'Chilli': {
'description': 'Important spice crop with high value',
'precautions': [
'Requires well-drained fertile soil',
'Irrigate carefully to avoid flower drop',
'Control fruit borer and mites',
'Harvest at color break stage'
]
},
'Pepper': {
'description': 'Black pepper, important spice crop',
'precautions': [
'Plant in well-drained red loamy soil',
'Provide support with standards or trellis',
'Control quick wilt and pollu beetle',
'Harvest when berries turn orange-red',
'Provide shade during initial growth'
]
},
'Turmeric': {
'description': 'Important spice crop with medicinal value',
'precautions': [
'Plant in well-drained fertile soil',
'Treat seed rhizomes with fungicide',
'Control leaf spot diseases',
'Harvest after 8-9 months when leaves dry'
]
},
'Tobacco': {
'description': 'Commercial crop mainly for export',
'precautions': [
'Requires well-drained sandy loam soils',
'Needs careful curing after harvest',
'Follow government regulations',
'Practice crop rotation'
]
},
'Sweet Potato': {
'description': 'Nutritious root vegetable rich in vitamins',
'precautions': [
'Plant in well-drained sandy loam soil',
'Use vine cuttings for propagation',
'Control sweet potato weevil',
'Harvest when leaves turn yellow',
'Cure properly before storage'
]
},
'Mango': {
'description': 'Important fruit crop of Andhra Pradesh',
'precautions': [
'Plant in well-drained deep soils',
'Prune for proper canopy management',
'Control mango hopper and fruit fly',
'Harvest at proper maturity'
]
},
'Banana': {
'description': 'Important fruit crop with high yield',
'precautions': [
'Requires heavy irrigation and fertilization',
'Plant disease-free tissue culture plants',
'Control sigatoka leaf spot disease',
'Support plants during fruiting'
]
},
'Coconut': {
'description': 'Important plantation crop of coastal areas',
'precautions': [
'Plant in coastal sandy soils',
'Apply balanced fertilizers regularly',
'Control rhinoceros beetle',
'Intercrop with cocoa or pepper'
]
},
'Cashew': {
'description': 'Important plantation crop for export',
'precautions': [
'Plant in well-drained sandy soils',
'Prune for proper shape',
'Control tea mosquito bug',
'Harvest nuts when apple turns pink'
]
},
'Soybean': {
'description': 'Oilseed crop rich in protein',
'precautions': [
'Plant in well-drained soils',
'Inoculate seeds with rhizobium',
'Control yellow mosaic virus',
'Harvest when leaves yellow and drop'
]
},
'Sunflower': {
'description': 'Important oilseed crop',
'precautions': [
'Plant in well-drained soils',
'Provide support if needed',
'Control head borer',
'Harvest when back of head turns yellow'
]
},
'Jowar (Sorghum)': {
'description': 'Traditional millet crop',
'precautions': [
'Drought resistant crop',
'Control shoot fly in early stages',
'Harvest when grains are hard'
]
},
'Bajra (Pearl Millet)': {
'description': 'Traditional drought-resistant crop',
'precautions': [
'Grows well in poor soils',
'Control downy mildew',
'Harvest when grains are hard'
]
}
}
# District-wise climate information (unchanged)
district_climate = {
'Anantapur': {
'description': 'Hot and dry climate with low rainfall',
'soil': 'Red sandy loam soils',
'avg_temp': '28-40°C',
'avg_rainfall': '500-600 mm'
},
'Chittoor': {
'description': 'Moderate climate with some hilly areas',
'soil': 'Red soils and black cotton soils',
'avg_temp': '22-38°C',
'avg_rainfall': '900-1000 mm'
},
'East Godavari': {
'description': 'Coastal district with high humidity',
'soil': 'Alluvial and deltaic soils',
'avg_temp': '24-36°C',
'avg_rainfall': '1000-1100 mm'
},
'Guntur': {
'description': 'Coastal plains with hot climate',
'soil': 'Black cotton soils',
'avg_temp': '25-38°C',
'avg_rainfall': '800-900 mm'
},
'Krishna': {
'description': 'Coastal district with fertile delta',
'soil': 'Alluvial and black soils',
'avg_temp': '24-36°C',
'avg_rainfall': '900-1000 mm'
},
'Kurnool': {
'description': 'Semi-arid climate with low rainfall',
'soil': 'Red soils and black soils',
'avg_temp': '26-40°C',
'avg_rainfall': '600-700 mm'
},
'Nellore': {
'description': 'Coastal district with moderate rainfall',
'soil': 'Red soils and sandy loams',
'avg_temp': '24-36°C',
'avg_rainfall': '1000-1100 mm'
},
'Prakasam': {
'description': 'Mixed coastal and dry climate',
'soil': 'Red soils and sandy loams',
'avg_temp': '25-38°C',
'avg_rainfall': '800-900 mm'
},
'Srikakulam': {
'description': 'Northern coastal district with good rainfall',
'soil': 'Red and alluvial soils',
'avg_temp': '22-34°C',
'avg_rainfall': '1100-1200 mm'
},
'Visakhapatnam': {
'description': 'Coastal district with hilly terrain',
'soil': 'Red and laterite soils',
'avg_temp': '22-33°C',
'avg_rainfall': '1000-1100 mm'
},
'Vizianagaram': {
'description': 'Coastal district with moderate climate',
'soil': 'Red and alluvial soils',
'avg_temp': '23-35°C',
'avg_rainfall': '1000-1100 mm'
},
'West Godavari': {
'description': 'Fertile delta region with high humidity',
'soil': 'Alluvial and black soils',
'avg_temp': '24-36°C',
'avg_rainfall': '1000-1100 mm'
},
'YSR Kadapa': {
'description': 'Hot and dry climate with low rainfall',
'soil': 'Red soils and black soils',
'avg_temp': '27-40°C',
'avg_rainfall': '600-700 mm'
}
}
# Prediction function (unchanged)
def predict_crop(district, month, crop_choice=None):
# Get current temperature and rainfall based on district and month
temp = df[(df['District'] == district) & (df['Month'] == month)]['Temperature'].mean()
rainfall = df[(df['District'] == district) & (df['Month'] == month)]['Rainfall'].mean()
soil_type = df[df['District'] == district]['Soil_Type'].mode()[0]
# Prepare input for model
input_data = {
'District': district,
'Month': month,
'Temperature': temp,
'Rainfall': rainfall,
'Soil_Type': soil_type
}
# If user has selected a crop
if crop_choice and crop_choice != "I don't know":
input_data['Crop'] = crop_choice
input_df = pd.DataFrame([input_data])
input_encoded = pd.get_dummies(input_df, columns=['District', 'Month', 'Soil_Type', 'Crop'])
# Ensure all columns are present (add missing with 0)
train_columns = pd.get_dummies(df, columns=['District', 'Month', 'Soil_Type', 'Crop']).columns.drop('Suitability')
for col in train_columns:
if col not in input_encoded.columns:
input_encoded[col] = 0
input_encoded = input_encoded[train_columns]
prediction = model.predict(input_encoded)[0]
if prediction == 1:
result = f"✅ {crop_choice} is suitable to grow in {district} during {month}."
precautions = crop_info[crop_choice]['precautions']
precautions_text = "\n".join([f"• {precaution}" for precaution in precautions])
output = f"{result}\n\n📌 Precautions:\n{precautions_text}"
else:
alternatives = get_alternative_crops(district, month)
alt_text = "\n".join([f"• {crop}" for crop in alternatives[:3]])
output = f"❌ {crop_choice} is not recommended for {district} in {month}.\n\n🌱 Better alternatives:\n{alt_text}"
else:
# Recommend best crops
recommended_crops = get_alternative_crops(district, month)
rec_text = "\n".join([f"• {crop}" for crop in recommended_crops[:5]])
# Get climate info
climate = district_climate[district]
climate_text = (
f"🌡️ Avg Temperature: {climate['avg_temp']}\n"
f"🌧️ Avg Rainfall: {climate['avg_rainfall']}\n"
f"🌱 Soil Type: {climate['soil']}"
)
output = (
f"🌾 Recommended crops for {district} in {month}:\n\n{rec_text}\n\n"
f"📌 District Climate Info:\n{climate_text}"
)
return output
def get_alternative_crops(district, month):
# Get current temperature and rainfall based on district and month
temp = df[(df['District'] == district) & (df['Month'] == month)]['Temperature'].mean()
rainfall = df[(df['District'] == district) & (df['Month'] == month)]['Rainfall'].mean()
soil_type = df[df['District'] == district]['Soil_Type'].mode()[0]
# Test all crops and get suitability scores
crop_scores = []
for crop in crops:
input_data = {
'District': district,
'Month': month,
'Temperature': temp,
'Rainfall': rainfall,
'Soil_Type': soil_type,
'Crop': crop
}
input_df = pd.DataFrame([input_data])
input_encoded = pd.get_dummies(input_df, columns=['District', 'Month', 'Soil_Type', 'Crop'])
# Ensure all columns are present (add missing with 0)
train_columns = pd.get_dummies(df, columns=['District', 'Month', 'Soil_Type', 'Crop']).columns.drop('Suitability')
for col in train_columns:
if col not in input_encoded.columns:
input_encoded[col] = 0
input_encoded = input_encoded[train_columns]
# Get probability instead of binary prediction
proba = model.predict_proba(input_encoded)[0][1]
crop_scores.append((crop, proba))
# Sort by probability
crop_scores.sort(key=lambda x: x[1], reverse=True)
return [crop for crop, score in crop_scores if score > 0.7]
# Custom CSS for styling (unchanged)
css = """
.gradio-container {
font-family: 'Poppins', sans-serif;
background: linear-gradient(135deg, #f5f7fa 0%, #c3cfe2 100%);
}
.title {
text-align: center;
color: #2c3e50;
font-size: 28px;
font-weight: 600;
margin-bottom: 20px;
background: linear-gradient(90deg, #4b6cb7 0%, #182848 100%);
-webkit-background-clip: text;
-webkit-text-fill-color: transparent;
}
.description {
text-align: center;
color: #4a5568;
margin-bottom: 30px;
font-size: 16px;
}
.input-section {
background: white;
padding: 20px;
border-radius: 10px;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
margin-bottom: 20px;
}
.input-label {
font-weight: 500;
color: #2d3748;
margin-bottom: 8px;
display: block;
}
.output-section {
background: white;
padding: 25px;
border-radius: 10px;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
min-height: 200px;
font-size: 16px;
line-height: 1.6;
white-space: pre-wrap;
}
.output-title {
color: #2c3e50;
font-weight: 600;
margin-bottom: 15px;
font-size: 20px;
border-bottom: 2px solid #e2e8f0;
padding-bottom: 8px;
}
.btn-primary {
background: linear-gradient(90deg, #4b6cb7 0%, #182848 100%);
border: none;
color: white;
padding: 12px 24px;
border-radius: 8px;
font-weight: 500;
cursor: pointer;
transition: all 0.3s ease;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
}
.btn-primary:hover {
transform: translateY(-2px);
box-shadow: 0 7px 14px rgba(0, 0, 0, 0.1);
}
.select-dropdown, .text-input {
width: 100%;
padding: 12px;
border: 1px solid #e2e8f0;
border-radius: 8px;
font-size: 16px;
transition: all 0.3s ease;
}
.select-dropdown:focus, .text-input:focus {
border-color: #4b6cb7;
box-shadow: 0 0 0 3px rgba(75, 108, 183, 0.2);
outline: none;
}
.footer {
text-align: center;
margin-top: 30px;
color: #718096;
font-size: 14px;
}
.success {
color: #2e7d32;
}
.warning {
color: #d32f2f;
}
.recommendation {
background: #f0f4f8;
padding: 15px;
border-radius: 8px;
margin-top: 15px;
border-left: 4px solid #4b6cb7;
}
.crop-image {
max-width: 100%;
border-radius: 8px;
margin-top: 15px;
box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
}
"""
# Gradio Interface
with gr.Blocks(css=css) as demo:
gr.Markdown("""
🌱 Applications of Machine-Learning District-wise Climate,Soil & Water Based Crop Detection for Sustainable Cropping
""")
with gr.Row():
with gr.Column():
with gr.Group(visible=True) as input_section:
gr.Markdown("### 📍 Enter Your Farming Details")
district = gr.Dropdown(
label="Select Your District",
choices=districts,
value="Krishna",
interactive=True,
elem_classes="select-dropdown"
)
month = gr.Dropdown(
label="Select Planting Month",
choices=months,
value=datetime.now().strftime("%B"),
interactive=True,
elem_classes="select-dropdown"
)
crop_choice = gr.Dropdown(
label="Do you have a specific crop in mind? (Select 'I don't know' for recommendations)",
choices=["I don't know"] + sorted(crops),
value="I don't know",
interactive=True,
elem_classes="select-dropdown"
)
submit_btn = gr.Button("Get Recommendation", variant="primary", elem_classes="btn-primary")
with gr.Column():
with gr.Group(visible=True) as output_section:
# Remove the title from the output section
output = gr.Textbox(
label="",
interactive=False,
lines=15,
elem_classes="output-section"
)
gr.Markdown("""
""")
submit_btn.click(
fn=predict_crop,
inputs=[district, month, crop_choice],
outputs=output
)
# Launch the app
if __name__ == "__main__":
demo.launch()