🌊 AI-Aquatica

AI-Aquatica is a comprehensive open-source Python library designed to analyze water quality data using advanced AI and statistical tools.
It facilitates preprocessing, modeling, visualization, and reporting of hydrochemical datasets with minimal effort – empowering researchers and professionals in hydrology, ecology, and environmental monitoring.

---

✨ Features

• ✅ Data Import: Load datasets from CSV, Excel, JSON, SQL, NoSQL, and APIs.
• 🧼 Data Cleaning: Remove duplicates and handle missing values via multiple strategies.
• 📏 Data Standardization: Normalize and standardize data (Z-score, MinMax, log, sqrt, Box-Cox).
• 🧠 Missing Data Imputation: Fill gaps with:
  • Mean, Median, Mode
  • KNN Imputer
  • Regression Imputer
  • Autoencoder Neural Network
• ⚖️ Ion Balance: Detect chemical inconsistencies and auto-correct based on ionic ratios.
• 📊 Statistical Analysis: Get descriptive statistics, correlation matrices, ANOVA, time series decomposition.
• 🤖 AI/ML Modeling:
  • Regression & Classification (Logistic, SVM, Tree, RF)
  • Clustering (KMeans, DBSCAN)
  • Anomaly Detection (LOF, Isolation Forest)
  • Synthetic Data (GAN-based generation)
• 📈 Visualization:
  • Basic: Line, Bar, Pie, Scatter, Heatmaps
  • Advanced: PCA, t-SNE, Interactive Bubble Charts
• 📝 Report Generation:
  • Automatic HTML reports (statistics, ML evaluation, recommendations)

---

🛠 Installation

pip install ai-aquatica

Optional extras provide deep-learning and interactive visualization support:

# Install TensorFlow-powered utilities
pip install "ai-aquatica[deep_learning]"

# Install Plotly-based interactive charts
pip install "ai-aquatica[interactive]"

# Or grab everything
pip install "ai-aquatica[all]"

Or from GitHub:

git clone https://github.com/TyMill/AI-Aquatica.git
cd AI-Aquatica
pip install -e .[all]

Full guide: installation.md

---

📘 Documentation

Read the full documentation on GitHub Pages:
👉 https://tymill.github.io/AI-Aquatica/

Explore individual usage examples:

• usage_data_cleaning.md
• usage_data_loading.md
• usage_missing_data.md
• usage_statistical_analysis.md
• ... and more!

---

💡 Quick Start Example

from ai_aquatica.ml_analysis import train_classification_model
import pandas as pd
import numpy as np

# Create mock dataset
df = pd.DataFrame({
    'NO3': np.random.rand(100),
    'pH': np.random.rand(100),
    'target': np.random.randint(0, 2, 100)
})

X = df[['NO3', 'pH']]
y = df['target']

# Train a random forest classification model using AI-Aquatica's helper
model = train_classification_model(X, y, model_type='random_forest')
print("Model trained successfully.")

Report Generation Templates

AI-Aquatica ships with ready-to-use Jinja2 templates stored in ai_aquatica/templates. By default the report utilities render these files to create:

• statistical_report.html (plus an accompanying heatmap.png chart),
• interpretation_report.html,
• further_analysis_report.html.

The templates are available immediately after installation, but if you installed only the minimal dependencies make sure jinja2 is present:

pip install jinja2

You can point the report functions to your own template directory by passing the template_dir parameter. The directory should contain files named like the bundled templates so the engine can find them.

from ai_aquatica.report_generation import generate_statistical_report

custom_templates = "/path/to/my/templates"  # folder with statistical_report_template.html, etc.
generate_statistical_report(
    data=df,
    report_path="reports/wq_stat_report.html",
    template_dir=custom_templates,
)

Need more control? Copy the files from ai_aquatica/templates, modify them, and point template_dir to the folder containing your customized versions.

---

🤝 Contributing

Contributions are welcome! Please feel free to fork the repo and submit a pull request.
We especially welcome:

• New preprocessing or AI models
• Example notebooks / visual dashboards
• Dataset integrations

---

📄 License

This project is licensed under the MIT License.

---

🙏 Acknowledgments

Special thanks to:

• Open-source contributors
• Environmental data science community
• University of Szczecin & BNP Paribas for ongoing support

---

📫 Questions? Suggestions? Open an issue or email the maintainer.