BeeUnity - Predictive Analytics for Community Beekeeping

A deep learning-powered WordPress plugin for predictive analytics in community beekeeping in Makueni County, Kenya.

Features

🤖 Deep Learning Models

• Colonization Prediction: Predicts optimal times for hive colonization based on weather and vegetation data
• Disease & Pest Detection: Early detection of diseases (American Foulbrood, Nosema) and pests (Varroa mites, Small Hive Beetle, Wax Moth)
• Yield Optimization: Predicts honey yield and provides optimization recommendations

📊 Data Collection & Analysis

• Weather Data: Real-time weather data from Open-Meteo API (temperature, precipitation, humidity, wind, solar radiation)
• Vegetation Data: NDVI and flowering indices from Google Earth Engine (Sentinel-2/MODIS)
• Automated Sync: Daily automated data synchronization
• Data Processing: Cleaning, merging, interpolation, and feature engineering

📈 Advanced Visualizations

• Interactive maps with hive locations (Leaflet.js)
• Time series charts (Chart.js)
• Heatmaps for risk visualization
• Bounding box visualizations
• Scatter plots and trend analysis
• Spatial and map visuals
• Risk and prediction dashboards
• Comparative hive analysis

🌍 Multi-Language Support

• English
• Kiswahili (Swahili)
• Kamba (Local Kenyan language)

🛒 Marketplace Integration

• Connect farmers to honey buyers
• Product listings (honey, beeswax, propolis)
• Farmer-to-market connections

Installation

1. Upload Plugin

   Upload the 'Matthew@B' folder to /wp-content/plugins/

2. Activate Plugin

   Activate BeeUnity through the WordPress plugins menu

3. Install Python Dependencies (for ML training)

   cd wp-content/plugins/Matthew@B/ml-scripts
   pip install -r requirements.txt

4. Database Setup

   • The plugin automatically creates necessary database tables on activation
   • Tables include: weather_data, ndvi_data, hive_observations, predictions, marketplace, training_data

5. Initial Data Sync

   • Go to BeeUnity Dashboard
   • Click "Sync Data Now" to fetch historical data from APIs

Configuration

API Keys

• Google Earth Engine API: Pre-configured in the plugin
• Open-Meteo: No API key required (free tier)

Hive Locations

The plugin is pre-configured with 9 hive locations in Makueni County:

• Hive 1: -1.772350, 37.592544
• Hive 2: -1.773990, 37.591450
• Hive 3: -1.774491, 37.591403
• Hive 4: -1.774371, 37.591403
• Hive 5: -1.774246, 37.591352
• Hive 6: -1.773918, 37.591284
• Hive 7: -1.773954, 37.591218
• Hive 8: -1.773833, 37.591188
• Hive 9: -1.773732, 37.591146

Data Collection Period

• Start Date: 2023-01-01
• End Date: 2025-12-31

Usage

Dashboard

Access the main dashboard at: WordPress Admin > BeeUnity > Dashboard

Features:

• Hive location map
• Weather overview
• Quick actions (Sync Data, Train Models)
• Alerts and notifications

Colonization Predictions

1. Go to BeeUnity > Predictions
2. Select a hive
3. Click "Run Colonization Prediction"
4. View optimal colonization windows with confidence scores

Disease Detection

1. Go to BeeUnity > Disease & Pests
2. Select a hive
3. Click "Detect Risks"
4. Review risk levels and recommendations

Yield Optimization

1. Go to BeeUnity > Yield Optimizer
2. Select a hive
3. Set harvest date
4. View predicted yield and optimization recommendations

Marketplace

1. Farmers can create listings at BeeUnity > Marketplace
2. Use shortcode [beeunity_marketplace] on any page
3. Buyers can browse and contact sellers

Shortcodes

Hive Status

[beeunity_hive_status hive_id="1"]

Predictions

[beeunity_predictions hive_id="1" type="colonization"]

Weather Widget

[beeunity_weather location="Makueni"]

Marketplace

[beeunity_marketplace view="grid"]

Deep Learning Models

Training Models

To train the deep learning models with collected data:

1. Navigate to ML scripts directory

   cd wp-content/plugins/Matthew@B/ml-scripts

2. Train Colonization Model

   python train_colonization_model.py ../../../uploads/beeunity/models/colonization_training_data.json

3. Train Disease Detection Model

   python train_disease_model.py ../../../uploads/beeunity/models/disease_training_data.json

4. Train Yield Prediction Model

   python train_yield_model.py ../../../uploads/beeunity/models/yield_training_data.json

Model Architecture

• Colonization Model: Deep neural network with 128-64-32-16-1 neurons
• Disease Detection: Multi-output model for disease and pest classification
• Yield Prediction: Regression model with 256-128-64-32-16-1 neurons

All models use:

• ReLU activation functions
• Batch normalization
• Dropout for regularization
• Adam optimizer
• Early stopping

Database Schema

Weather Data

• hive_id, date_recorded, temperature_max, temperature_min, temperature_avg
• precipitation, humidity, wind_speed, solar_radiation

NDVI Data

• hive_id, date_recorded, ndvi_value, evi_value, flowering_index

Hive Observations

• hive_id, observation_date, colonization_status, bee_population
• honey_yield, disease_detected, pest_detected, notes, images, audio

Predictions

• hive_id, prediction_date, prediction_type, prediction_value
• confidence_score, model_version

Marketplace

• farmer_id, product_type, quantity, price, description
• harvest_date, status

API Endpoints

Weather Data

• Source: Open-Meteo Archive API
• Endpoint: https://archive-api.open-meteo.com/v1/archive
• Data: Temperature, precipitation, humidity, wind speed, solar radiation

Vegetation Data

• Source: Google Earth Engine (MODIS/Sentinel-2)
• Data: NDVI, EVI, flowering index

Cron Jobs

The plugin schedules daily data synchronization:

• Hook: beeunity_daily_data_sync
• Frequency: Daily
• Function: Syncs weather and NDVI data for all hives

File Structure

Matthew@B/
├── beeunity.php (Main plugin file)
├── includes/
│   ├── class-beeunity-activator.php
│   ├── class-beeunity-deactivator.php
│   ├── class-beeunity-db.php
│   ├── class-beeunity-ajax.php
│   ├── class-beeunity-i18n.php
│   ├── data-collection/
│   │   ├── class-openmeteo-api.php
│   │   ├── class-gee-api.php
│   │   └── class-data-processor.php
│   ├── ml/
│   │   ├── class-ml-engine.php
│   │   ├── class-colonization-predictor.php
│   │   ├── class-disease-detector.php
│   │   └── class-yield-optimizer.php
│   ├── marketplace/
│   │   └── class-marketplace.php
│   └── visualizations/
├── admin/
│   ├── class-beeunity-admin.php
│   ├── views/
│   ├── css/
│   └── js/
├── public/
│   ├── class-beeunity-public.php
│   ├── css/
│   └── js/
├── ml-scripts/
│   ├── train_colonization_model.py
│   ├── train_disease_model.py
│   ├── train_yield_model.py
│   └── requirements.txt
└── languages/

Requirements

WordPress

• WordPress 5.8+
• PHP 7.4+
• MySQL 5.7+

Python (for ML training)

• Python 3.8+
• TensorFlow 2.13+
• Scikit-learn 1.3+
• NumPy, Pandas, Matplotlib

JavaScript Libraries (CDN)

• Chart.js 4.4.0
• Leaflet.js 1.9.4
• TensorFlow.js 4.11.0

Support

For support and questions:

• Email: support@beeunity.makueni.org
• Documentation: https://docs.beeunity.makueni.org

License

GPL v2 or later

Credits

• Developer: The Open University Of Kenya,Data Science (Cohort2)- Group 2.
• Data Sources: Open-Meteo, Google Earth Engine
• Location: Makueni County, Kenya

Version History

1.0.0 (2026-01-12)

• Initial release
• Colonization prediction
• Disease and pest detection
• Yield optimization
• Marketplace integration
• Multi-language support (English, Swahili, Kamba)
• 9 hive locations in Makueni County
• Historical data: 2023-2025